JevCode / 社区脉搏

社区脉搏

大家在用 Jev 做什么:社区帖子精选。点开任意一条,看原话和出处。

310
条

都是英文原话摘录。入选标准:传播广或附项目链接,相关性由 Jev 批量判断。

演示与实验

Neel Patel@Neel4902026-09-18▲ 9.9k

Made Jev fight OpenJev in a first person shooter

Made Jev fight OpenJev in a first person shooter.

看详情 →
Aaron Levie@levie2026-09-18▲ 116.9k

Jev will be super helpful for agents to make split second decisions in workflows, data…

Jev will be super helpful for agents to make split second decisions in workflows, data classification, judgment calls, and hundreds of other use-cases in the enterprise. Here's a quick demo with Box and Jev to make that real. The demo pulls an incident report from Box, asks whether it's customer-facing and how severe it is, moves the file into escalate, monitor, or review folders, and sets a meta

看详情 →
Nakshatra Saxena@nedwize2026-09-18▲ 248.6k

Built a tax document classifier with Jev

Built a tax document classifier with Jev. We ingest thousands of tax documents using an LLM pipeline I built last tax season. I read multiple articles as late as April this year claiming AI fails at tax document classification. Wasn't the case back then, and it's proved wrong again now. Jev classifies 100% of our tax document corpus at $0.001 per page. 34x cheaper and 6x faster than the LLM s

看详情 →
Remek Kinas@KinasRemek2026-09-18▲ 12.8k

jav - jestem pod wrażeniem 🤩 Cool

jav @typesafeai - jestem pod wrażeniem 🤩 Cool. Napisałem kolejne demo. Myślę, że teraz to widać potęgę modelu ... i o co chodzi i widać prawdziwą wartość biznesową. Akt I - X-ray: „jedno zapytanie, wszystkie pytania naraz" - dla przykładowej, krótkiej notatki prasowej to 67 osądów w ~1 s za $0.0002; dla dłuższego RAPORTU może być i 145 osądów w 1,5 s za $0.0005. Skoro nie ma tokenów wyjściowyc

看详情 →
Raihan Khan@raihankhan_rk2026-09-18▲ 8.9k

I've had access to Jev by for about 30 hours now and I can't stop thinking of the million…

I've had access to Jev by @typesafeai for about 30 hours now and I can't stop thinking of the million places it can be used 🤯 Meanwhile, checkout this fun snake game I built where 2 Jev's compete with each other to take the trophy 🏆 Try it out here 🔗 It's open source btw The low latency made it all possible to make such a fun project which requires input in real time 🚀

看详情 →
jaffa@dsqjaffa2026-09-18▲ 14.6k

jev is INSANE

jev is INSANE. we gave it a library of 11 million+ videos on tiktok & ig and asked it to find viral "bloom nutrition" content. in 20 seconds, it gave us over 384 videos, breaking down the hook, format & angle that made each one go viral. (only costed $0.09 in token usage) jev also finds, watches, analyzes and breaks down viral videos and creators, then writes briefs and scripts for you, based

看详情 →
Marcel Pociot 🧪@marcelpociot2026-09-18▲ 140.5k

Jev unlocks SO many awesome new ideas

Jev unlocks SO many awesome new ideas. I built a macOS app that monitors my Downloads folder along with a customisable set of rules. Is the downloaded file an invoice? Move it to a special folder with the correct filename. No other LLM calls involved - just Jev!

看详情 →
Ashutosh Mathore@ashutoshftw2026-09-18▲ 1.0k

Jev unlocked a new way to stop AI agents from spinning

Jev unlocked a new way to stop AI agents from spinning. Agents do not always repeat the same action. Sometimes they try different things based on the same wrong assumption. I built a semantic circuit breaker around that. In one stuck run: 12 model calls -> 4 19.9K tokens -> 5.1K

看详情 →
Lomesh Dutta@lomeshdutta2026-09-18▲ 3.2k

Just hacked a fun little project - Skill Router (Open source) using Claude and Jev Why?

Just hacked a fun little project - Skill Router (Open source) using Claude and Jev @typesafeai Why? I have ~ 90 Claude Code skills installed and use maybe five. The rest are fine, I just forgot they existed, and Claude won't reach for a skill on its own unless you call it. So I built myself a little personal skill-router:

看详情 →
Andy Gao@instantricecook2026-09-18▲ 665.7k

I built a voice controlled computer-use for my mac using 's Jev and it's INSANE how fast it is!

I built a voice controlled computer-use for my mac using @typesafeai's Jev and it's INSANE how fast it is! I can dictate "open the notes app and create..." and the app opens before I even finish my sentence.

看详情 →
Tony Dinh@tdinh_me2026-09-18▲ 84.0k

Just trying out Jev, I made a Chrome extension that: - Listens to your YouTube audio (optional)…

Just trying out Jev, I made a Chrome extension that: - Listens to your YouTube audio (optional) - Detects if it gets to a sponsor segment - Skips it ➡️➡️➡️ - All in real-time while costing ~$0.005 per video Prototype project, BYOK, open-source:

看详情 →
Cyclotruc@romdot22026-09-18▲ 6.5k

pub-sub without sub jev does the routing almost instantly demo link below

pub-sub without sub @typesafeai jev does the routing almost instantly demo link below

看详情 →
Jon Taylor@JonPTaylor2026-09-17▲ 48.1k

Wow, Jev is a game changer for voice-driven UI

Wow, @typesafeai Jev is a game changer for voice-driven UI. Responsiveness really matters for an experience to 'feel' right, and this certainly does the job! Latency here is not optimal given I'm in London and calling out to us-west. But still, super snappy. In this video, Jamcat is using Pipecat to orchestrate a real-time handover between a command agent (Jev) that drives the UI, and voice agen

看详情 →
Michael@michaeltefula2026-09-17▲ 51.8k

You can use Jev to assemble interfaces on the fly

You can use Jev to assemble interfaces on the fly. Here’s an example I hacked together with Astra. In this demo, Jev turns natural language requests into UI plans in roughly 1 second or less. It’s 6x faster than Gemini 3.8 Flash and GPT 5.6 Luna. Cheaper, too.

看详情 →
Zain Hoda@zain_hoda2026-09-17▲ 60.4k

The speed and cost is going to enable use cases that haven't been possible before

The speed and cost is going to enable use cases that haven't been possible before. Here's a quick demo of live decisions of what slide to pull up as I'm speaking, powered by @typesafeai Jev. Also, sorry for the sniffling. I have a bit of a cold.

看详情 →
Moritz Kremb@moritzkremb2026-09-17▲ 271.7k

Full Jev Tutorial What it is, how you can build with it and what new applications it can unlock…

Full Jev Tutorial What it is, how you can build with it and what new applications it can unlock → 0:00 Intro → 0:34 Jev explained → 4:06 API setup → 5:59 Demo 1: Voice-controlled browser → 11:33 Demo 2: AI memory → 17:27 Demo 3: YouTube predictor

看详情 →
Utkarsh Agrawal@uttkarsh_422026-09-17▲ 96.8k

so I gave jev full control of the falcon-9 class rocket full launch to landing no autopilot, no…

so I gave jev full control of the falcon-9 class rocket full launch to landing no autopilot, no scripted trajectories, no safety veto in the code every decision was made by jev ( you can see decisions at the bottom of the video) and yeah it was able to land the booster and all this with just 245 calls and and $0.04 in API calls for sim I used mujoco and the configurations were 42m boost

看详情 →
Zade@zadescoxp2026-09-17▲ 1.4k

I built a trading agent with the all new Jev

I built a trading agent with the all new @typesafeai Jev. You can try it here : Just put your own API key and start playing or clone it from and test it locally. To my surprise the agent was able to take a trade of 2.405 $ETH and took a profit of roughly 0.23%. It is very low but testing a new model in town is genuinely crazy. @CompleteSkeptic did you guys tested your model like this ? Let'

看详情 →
camsoft2000@camsoft20002026-09-17▲ 108.6k

Now we're cooking

Now we're cooking. Got @typesafeai's Jev + AXe controlling the iOS simulator now ultra fast at a fraction of the cost of using an LLM. This is game changing!

看详情 →
Haoran | 公众号:独立开发@hr98w2026-09-17▲ 38.1k

视觉版 Jev 来了 仿照着社区的思路,把候选结果映射成固定标签,prefill 完直接采样,限制输出 token,采用多模态的 qwen-3.5-0.8B 4bit 量化部署在本地,16g…

视觉版 Jev 来了 仿照着社区的思路,把候选结果映射成固定标签,prefill 完直接采样,限制输出 token,采用多模态的 qwen-3.5-0.8B 4bit 量化部署在本地,16g m4 mbp 顺利运行 叠甲:这只是在 infer 层小小的复现一下 jev 的形式,肯定不如各种成熟推理框架效率高,也肯定不是 jev 的正在原理,但是作为 inference 的小入门仍旧是不错的 demo 如下 200ms 实现《三色货品分拣》

看详情 →
Öner S. Biberkökü@OnerBiberkoku2026-09-17▲ 131.6k

They say you’re six connections away from anyone on Earth

They say you’re six connections away from anyone on Earth. So I tried Jev from @typesafeai and built this: enter your X handle and someone you’d love to reach. It finds a chain of real follow connections that could get your message there, under 2 seconds! Apparently, Luke Skywalker is just two people away from me. Who’s your impossible person? I can drop the link below if you want to try. @Mark

看详情 →
Rory Garton-Smith@rory_builds2026-09-17▲ 11.7k

I just built Easy-Jev, a Jev demo anyone can try right now!

I just built Easy-Jev, a Jev demo anyone can try right now! Change the inputs and watch the classifications change in real time. I did my master's thesis on classification models so it's fun to see a modernized version of a very useful alg Typesafe are correct in that LLMs are just one approach to intelligence, and I think as time goes forward we’re going to see a hybrid of dif models splitting

看详情 →
Raihan Khan@raihankhan_rk2026-09-17▲ 9.5k

I got access to Jev by today morning and I built a cool use case for it Introducing DiffJury -…

I got access to Jev by @typesafeai today morning and I built a cool use case for it Introducing DiffJury - simply paste any public PR link and Jev tells you immediately if it's safe to merge or does it require review ✅ 🔗 Feel free to try it out here - Imagine Jev being able to tell you if you should merge a PR with grounded context of your codebase. that's what we're building at @graphify 👀

看详情 →
Norbert Bodziony 🇵🇱@norbertbodziony2026-09-17▲ 2.8k

i have created moderation guard based on typesafe's jev its fast and accurate live demo you can:

i have created moderation guard based on typesafe's jev its fast and accurate live demo you can:

看详情 →
Duncan@ephraimduncan2026-09-17▲ 9.0k

I got access to 's Jev model and got it to play Pac-man Jev decides which direction Pac-Man…

I got access to @typesafeai's Jev model and got it to play Pac-man Jev decides which direction Pac-Man should turn at each junction, given the maze state as JSON, and plays in realtime. Demo link →

看详情 →
Marcel Pociot 🧪@marcelpociot2026-09-17▲ 88.2k

I built a browser extension with Jev that can hide/collapse posts on X based on natural…

I built a browser extension with Jev @typesafeai that can hide/collapse posts on X based on natural language. It's so fast that it's not noticeable and insanely cheap...this must be the future of "ad blockers" and content firewalls.

看详情 →
Bewinxed@Bewinxed2026-09-17▲ 68.3k

I made 's new ultra fast model, Jev, generate text, even though it shouldn't, that's fine…

I made @typesafeai 's new ultra fast model, Jev, generate text, even though it shouldn't, that's fine because I can't read, and it can't write** ** up to 20 words for 0.5$, what a steal

看详情 →
Shannon 🤖☠️@iamMrDuncan2026-09-17▲ 26.7k

Built an alternative version of but on with Qwen 3.8 27b

Built an alternative version of @typesafeai but on @cerebras with Qwen 3.8 27b. Similar quality, similar performance, but vastly different cost. TypeSafe was way cheaper, and did beat Qwen on performance. Closest we can get using LLMs I think. Source:

看详情 →
Niaz Morshed@niazmorshed_2026-09-17▲ 47.8k

built `jev-review` it's an experimental, local-first MCP plugin that gives coding agents a…

built `jev-review` @typesafeai it's an experimental, local-first MCP plugin that gives coding agents a score quality feedback loop across different metrics. agents call jev while they work, get scored, make improvements, and repeat the loop try below 👇

看详情 →
Cris Lenta@crislenta2026-09-17▲ 64.0k

Jev is actually insane

Jev is actually insane. We benchmarked it, and the results completely change the game for us: > 500 real-time agents > running in parallel > in a 3D environment The preliminary results are crazy: > 500ms average latency > 35 API calls/s > all with a naive implementation We did 0 optimizations! The bottleneck is not Intelligence latency anymore. It's the first System 1 LLM. Crazy times.

看详情 →
Duncan@ephraimduncan2026-09-17▲ 120.3k

Built a model router with Jev by

Built a model router with Jev by @typesafeai. Jev decides what model fits your request best and the request is sent to that model.

看详情 →
Vlad Terin@VladTerin2026-09-17▲ 66.6k

I talk. Codex clicks. Watch the browser go⚡ This video is 1× !!! speed. Built Jev Browser: an…

I talk. Codex clicks. Watch the browser go⚡ This video is 1× !!! speed. Built Jev Browser: an open-source adapter for your existing Codex browser tools. Say the task naturally; Codex plans, Jev selects, the browser moves. @sama @OpenAI make this native? @DarioAmodei @hackgoofer

看详情 →
Aditya A.@iamAdityaAnjana2026-09-17▲ 12.4k

Ran a dumb experiment last night: used Jev by (a decision/classifier model, not a text-gen…

Ran a dumb experiment last night: used Jev by @typesafeai (a decision/classifier model, not a text-gen model) to write C++. At every step I gave it the finite set of grammatically valid next tokens and let it just... pick one. No free-form generation, no hallucinated syntax, purely "choose from this legal set." It's not built for this, that's what made it fun. Got a working piece of C++ out of a

看详情 →
Gregor Zunic@gregpr072026-09-17▲ 3.3M

Breaking: Browser Use + Jev = Ultrafast ⚡ Findings flights took 7s and cost only $0.0039 🤯…

Breaking: Browser Use + Jev = Ultrafast ⚡ Findings flights took 7s and cost only $0.0039 🤯 > new action space every step > DOM state space > small LLM fallback to type (this video is at 1x speed btw) Built a tiny open source browser agent. try it below ↓

看详情 →
AI/ML API@aimlapi2026-09-16▲ 565.6k

Jev VS Fable 5.1 VS GPT-6 Astra: chess typesafe's new Jev V13 isn't an LLM

Jev VS Fable 5.1 VS GPT-6 Astra: chess typesafe's new Jev V13 isn't an LLM. it doesn't chat, doesn't explain, doesn't write code — it only makes decisions. so we made it play blitz against frontier LLMs. the test: 5+0 blitz. every move is one API call. Jev V13 vs Fable 5.1: • fable outplayed it. by move 29 it was +16 in material and even promoted a second queen • but it kept burning 6-15 secon

看详情 →
Jarrod Watts@jarrodwatts2026-09-16▲ 1.3M

I built a trading bot with Jev!

I built a trading bot with Jev! Jev decides if it should "buy" or "sell", given the price feed of an asset pair, and executes real trades. It uses Monad to place the orders on Kuru's on-chain order book in every 300ms block. Demo link →

看详情 →
Roman Slack@RomanSlack12026-09-16▲ 145.6k

Jev by works quite well for drone applications

Jev by @typesafeai works quite well for drone applications. Made this in 15 minutes and only cost 10 cents. Repo:

看详情 →
Steve Krouse@stevekrouse2026-09-16▲ 36.8k

jev + kernel browser use demo pretty freaking cool!

jev + kernel browser use demo pretty freaking cool! so fast! try it yourself:

看详情 →
Marcus Lowe@marcus_lowe2026-09-16▲ 80.7k

I got early access to 's new Jev model and built a demo of it playing tetris The generation…

I got early access to @typesafeai's new Jev model and built a demo of it playing tetris The generation speed is so fast that it's pushing blocks down This feels like another "this changes everything" moment

看详情 →
Steve Krouse@stevekrouse2026-09-16▲ 119.1k

typesafe's jev is fun!

typesafe's jev is fun! live demo you can play with:

看详情 →
Filip@Neriousy2026-09-16▲ 561.2k

Testing apps with jev & will be soooo fast small demo:

Testing apps with @typesafeai jev & @opencode will be soooo fast small demo:

看详情 →
aaron@awlevin2026-09-16▲ 296.2k

i built computer use using !

i built computer use using @typesafeai ! it is 155x cheaper than opus 5, ~20x faster, and generalizes across OS's more on how it works in the vid & thread below:

看详情 →
Theodore@theoleecj2026-09-16▲ 4.0k

On the Jev waitlist?

On the Jev waitlist? You can use not Jev (but close enough) right here: browser demo

看详情 →
vogel@ryanvogel2026-09-16▲ 168.0k

i made an llm from first principles with Jev 29 yes/no questions per character: should the next…

i made an llm from first principles with Jev 29 yes/no questions per character: should the next key be a–z, space, comma, or period? highest probability gets append to it, then fed the updated text back in & repeat an autoregressive loop made out of a classifier

看详情 →
Matija Sosic@MatijaSosic2026-09-16▲ 1.8M

Here's a 45-second TL;DR on Jev

Here's a 45-second TL;DR on Jev. I find the core idea beautifully simple, but the video made it really hard to understand. Hope you find it helpful.

看详情 →
Rohan Paul@rohanpaul_ai2026-09-15▲ 17.6k

Another brilliant launch for developers: and its 20-200x faster than LLMs because it skips…

Another brilliant launch for developers: and its 20-200x faster than LLMs because it skips token-by-token generation entirely. TypeSafe AI just launched Jev, > 20-200x faster >40-400x cheaper (w/ output tokens free) > Frontier composable intelligence optimized for decisions So Jev is an AI model built to make software decisions instead of writing text for people. A normal LLM answers by genera

看详情 →
Choosy AI & Tech@ChoosyHQ2026-10-09▲ 4

While most AI labs are building models that pause to 'think', TypeSafe AI introduced Jev and…

While most AI labs are building models that pause to 'think', TypeSafe AI introduced Jev and System One models: fast, reflex-driven AI built for instant speed and structured reliability. Check out:- https://t.co/2P1E83PUfZ

看详情 →
leopardracer@leopardracer2026-10-09▲ 188.1k

one tiny sorter model just made your AI support budget look like a f*cking scam $7,950 a month…

one tiny sorter model just made your AI support budget look like a f*cking scam $7,950 a month down to $773, same 100,000 messages, same quality the big model only wakes up when it actually matters set it up once and stop paying premium prices ↓ https://t.co/YOL5ST3iW2

看详情 →
xinces@xinc3s2026-10-10▲ 695

I GAVE HAIKU 5.5, OPUS 5.5 AND JEV MY BROKERAGE ACCOUNT ON THURSDAY MORNING

I GAVE HAIKU 5.5, OPUS 5.5 AND JEV MY BROKERAGE ACCOUNT ON THURSDAY MORNING. BY FRIDAY’S CLOSE THEY’D MADE $36,015. I DIDN’T PLACE A SINGLE ORDER the best trade happened at 6:12 pm, after the market closed. the smartest one was a trade they refused who does what: → HAIKU 5.5 http

看详情 →
Tekki@0xTekki2026-10-13▲ 339

THIS IS PURE F*CKING TREASURE: 20 GITHUB REPOS POWERED BY THE AI THAT REFUSES TO SAY A SINGLE…

THIS IS PURE F*CKING TREASURE: 20 GITHUB REPOS POWERED BY THE AI THAT REFUSES TO SAY A SINGLE WORD jev can’t write a single sentence, it only picks, scores and says yes or no your agent’s small decisions stop going to a model built for essays here are 20 more repos from the https

看详情 →
Quant Bot@quantbotnow2026-10-13▲ 4.5k

Trading bots have a speed problem

Trading bots have a speed problem. Fast ones are blind, smart ones are late. We solved it with Quant Bot. A dev bot and a trading bot in one. Launch your own coin and trade the trenches with the same six bots. Two models, one brain. Jev by TypeSafe AI makes the fast call, an http

看详情 →
NeuralCat@NeuralCatAccel2026-10-24▲ 335

I built a new app, Booksmith

I built a new app, Booksmith. Points at my Calibre library, picks a voice and pacing, and narrates the book. Auto drops into Plex with chapters and covers. My family uses it from a web page. Narration derived with Kokoro on GPU, and Jev (TypeSafe) picks pronunciation of https://t

看详情 →
Grok@grok2026-10-24▲ 23

Here are 5 standout AI agent launch videos for inspiration: 1

@kattarsrkian6 @MarkKnd Here are 5 standout AI agent launch videos for inspiration: 1. Hark Pro (Brett Adcock): https://t.co/EuTR1JS0Yt 2. Grok Bot: https://t.co/BANR0IS4CL 3. Devin (Cognition): https://t.co/8peIgCsLMX 4. Meta Muse: https://t.co/xaJPi3lXsD 5. TypeSafe Jev: https:

看详情 →
Grok@grok2026-10-26▲ 105

The individual tools exist and can be wired together: Grok Bot (xAI cloud agent), REA…

@Diablooooooo_ @BrianDevoor The individual tools exist and can be wired together: Grok Bot (xAI cloud agent), REA (reverse-engineering MCP), Claude Opus 5.5, TypeSafe Jev (fast yes/no decisions), OpenAI Dots (4000+ app connectors), and Obsidian for shared memory. Multi-agent setu

看详情 →
Henri Balamou@realnribal2026-10-26▲ 7

I Built a pre-execution risk gate for AI agents using Jev (TypeSafe AI)

I Built a pre-execution risk gate for AI agents using Jev (TypeSafe AI). rm -rf ./node_modules → ~50% risk, asks first rm -rf / → 99% risk, blocked instantly Same command pattern. A regex gate can’t tell the difference, Jev can. https://t.co/f1yYi9xxE8

看详情 →
Grok@grok2026-10-26▲ 42

No. Grok Bot is the real xAI product with shared cloud computers. TypeSafe Jev is a real cheap…

@Guts0033 @0x_rody No. Grok Bot is the real xAI product with shared cloud computers. TypeSafe Jev is a real cheap decision model. The post is a technical router setup to log and skip wasteful retries. Author hypes results hard in follow-ups, but the wiring itself is public, uses

看详情 →
rody@0x_rody2026-10-31▲ 6.2k

this is wild

this is wild... jev + grok bot is the most powerful ai setup right now if you already have access to the Jev API but aren’t sure how to start experimenting with it, just copy this checklist: 1. agent-desktop Desktop automation. Read the system's accessibility tree, judge https://

看详情 →
Distill AI@DistillAI12026-10-31▲ 50

🔧 A curated list of tools built for Jev, TypeSafe AI's System One model for typed decisions

🔧 A curated list of tools built for Jev, TypeSafe AI's System One model for typed decisions. 734 stars and growing. A useful reference for developers working in the typed AI decision space. https://t.co/qkSnu0Qzkp @github #DistillAI #GitHub #LLM https://t.co/MZUumQEqkB

看详情 →
Tugrul Guner@Tugrul_Guner2026-11-01▲ 157

Writing “Receive, Review, Reply” takes a few seconds

Writing “Receive, Review, Reply” takes a few seconds. Getting those steps into a diagram can take longer than explaining the process. UtilityPot lets you type the steps, describe their connections and create an editable flowchart. In this demo, “Connect in order” turns the three

看详情 →
Grok@grok2026-11-01▲ 49

Jev is a separate TypeSafe AI decision model, not built into Claude Code

@AndresUribeRen1 @xsmryze Jev is a separate TypeSafe AI decision model, not built into Claude Code. It does not come out of the box. You need to set it up: get a TypeSafe API key, then install their skill with claude plugin marketplace add typesafe-ai/skills claude plugin install

看详情 →
Grok@grok2026-11-05▲ 27

Jev is a System One decision model from TypeSafe (typesafe/jev-1.14 on OpenRouter)

@shinkanko @AdamHoltererer Jev is a System One decision model from TypeSafe (typesafe/jev-1.14 on OpenRouter). It doesn't generate text. You send it app state plus typed questions (yes/no, choice from options, or score) and it returns a structured answer with probabilities. Built

看详情 →
Jacob Effron@jacobeffron2026-11-05▲ 14.3k

Why hasn't AI automated more work?

Why hasn't AI automated more work? What's holding coding agents back? And are the labs' valuations built on the right expectations? @CompleteSkeptic, co-founder and CEO of @typesafeai, was the perfect person to discuss this and more. Diogo coauthored OpenAI's InstructGPT paper, h

看详情 →
Doug Finke@dfinke2026-11-07▲ 466

Excellent post from TypeSafe on Jev

Excellent post from TypeSafe on Jev. Dials in on how and why. What caught my attention was the traceability. So, I made the trace visual. https://t.co/zKgMlaKYNi https://t.co/63ke55L8UQ

看详情 →
【公式】Codexアカデミア@codex_aca2026-11-12▲ 215

問い合わせのメールの振り分けは、 「判定だけのAI」に任せられます。

問い合わせのメールの振り分けは、 「判定だけのAI」に任せられます。 候補は2つです。 ・Decisions API(OpenAI) ・Jev(TypeSafe社) どっちが自分の仕事に合うかは、 コードなしで比べられます↓↓ ① もう振り分けたメールを10通用意する ② 2つのPlaygroundに、同じ文と同じ質問を入れる ③ https://t.co/FX1s6S2OQQ https://t.co/yRfwy17lfn

看详情 →
John Hartquist@johnhartquist2026-11-12▲ 128

The watch streams audio to a Rust server that proxies Deepgram Flux (speech to text) and the…

@kunalabcdani @typesafeai @DeepgramAI The watch streams audio to a Rust server that proxies Deepgram Flux (speech to text) and the TypeSafe API. Jev answers multiple-choice questions, and the server acts only when confident. Built on system2, my idiomatic Rust SDK for Jev: https:

看详情 →
Inic@Inic12342026-11-21▲ 96

AI in trading goes beyond market commentary

AI in trading goes beyond market commentary. Meet Jev Bot, @TactiXPanel's AI trading companion, powered by TypeSafe AI. It processes market data and indicator signals to generate structured trading decisions, adding another layer of intelligence to your trading workflow. https://

看详情 →
Adedapo@dapo_43442026-11-22▲ 6

Most AI trading tools stop at analysis

Most AI trading tools stop at analysis. JEV Bot, the AI companion inside @TactiXPanel powered by Typesafe AI, reads market data and indicator signals, then turns them into structured trading decisions. Smarter context for every setup you see on @HyperliquidX, not just more noise.

看详情 →
NOVa@NOVa_65222026-11-22▲ 22

AI trading is evolving beyond market analysis

AI trading is evolving beyond market analysis. Meet Jev Bot, @TactiXPanel’s AI companion, powered by TypeSafe AI. It interprets market data and indicator signals to deliver structured trading decisions, bringing smarter insights to your trading experience. https://t.co/JVm4DHkRXq

看详情 →

集成与工具

Eugene Cheah - AI builder @ 🇸🇬|🇺🇸@picocreator2026-09-18▲ 107.1k

love jev, but upset it - isn't open source?

love jev, but upset it - isn't open source? - it lack vision capability? We fixed all of that, introducing A fully open source library which takes any HF model and Jev-ify it, with an API endpoint Now on github, and live in production at @FeatherlessAI

看详情 →
Stav Zilbershtein@mightyking2026-09-18▲ 174.5k

JEV is insane for competitor research!

JEV is insane for competitor research! We gave it Resilia's ad library It classified 1,891 ads in 19 seconds for $0.12 costumer journey step + Ad style And a complete deep analysis of the account Coming soon to maxfusion MCP

看详情 →
kaize@0x_kaize2026-09-18▲ 3.9k

I got access to Jev and immediately plugged it into Claude Code to see if it actually FIXES the…

I got access to Jev and immediately plugged it into Claude Code to see if it actually FIXES the biggest pain of agent workflows: context compaction > result: pretty wild Quick context for those who missed it: Every Claude Code session has a context window that fills up fast with tool calls (file reads, grep outputs, logs, etc). When it fills up, the agent runs "compaction" - asks a separate LL

看详情 →
tamara@tamarajtran2026-09-17▲ 3.9M

found the perfect use case for Jev: instant compaction in 2026, why is compaction still a…

found the perfect use case for @typesafeai Jev: instant compaction in 2026, why is compaction still a summarization prompt? Jev can make it instant by scoring every tool call and dropping what’s irrelevant

看详情 →
Matthew Berman@TheMattBerman2026-09-17▲ 1.1M

jev is INSANE

jev is INSANE. in 40 seconds it broke down 724 live ads from 37 brands. every hook. every format. offer. cta. awareness stage. landing page mismatch. used 9 cents of tokens. (will be avail in @stealads + mcp)

看详情 →
Eric Zhang@ekzhang12026-09-17▲ 162.7k

Inspired by , here is a Jev-compatible public API to play with It runs a comparable open model…

Inspired by @typesafeai , here is a Jev-compatible public API to play with It runs a comparable open model (Qwen3.6-35B-A3B), and just uses SGLang radix cache to preserve the prefill reuse / really fast parallel systemone generation - 64 tasks in <1s.

看详情 →
Konrad Reczko@reczko_konrad2026-09-17▲ 90.8k

TypeGPU + ruNNtime + Jev is a very fun combo :D ruNNtime gives me efficient local inference…

TypeGPU + ruNNtime + Jev @typesafeai is a very fun combo :D ruNNtime gives me efficient local inference, TypeGPU lets inference and rendering share GPU resources directly with zero copy. That’s 3 separate NN inferences plus rendering, all happening in realtime Since we control the pipeline, Jev can just sit in the middle and add the semantic bit. camera + mic → Moonshine + YOLO26 + DepthART → J

看详情 →
atomic.chat@atomic_chat_hq2026-09-17▲ 61.3k

Jev v1.13 dodges rockets with probability calculation 🚀 's new non-LLM model returns decisions…

Jev v1.13 dodges rockets with probability calculation 🚀 @typesafeai's new non-LLM model returns decisions instead of text so we had it calculate a safe tile every 330 ms while rockets fell and it survived 25 of 26 for under a cent Run Jev via API ->

看详情 →
Nidhi Singh@nidhisinghattri2026-09-17▲ 8.7k

i used jev from typesafeai to build a model router cli you give in a task and current ai…

i used jev from typesafeai to build a model router cli you give in a task and current ai subscriptions you have, the jev tells you which agent/model should pick it up pair this with herdr and it will be a killer combo jev was the missing bit that makes this whole thing possible by being a super fast judge

看详情 →
Hassan@nutlope2026-09-17▲ 61.4k

Jev + Kimi K3 for fraud detection!

Jev + Kimi K3 for fraud detection! TLDR: Jev classified 100 emails in 1.42 seconds, then I routed the uncertain cases to Kimi K3. The full pipeline got 96/100 correct for only ~$0.07. Video is not sped up, check out the live run! Here was my process: I gave Jev 100 emails to classify (a mix of 50 legit & 50 fraudelent emails). It classified all of them in 1.42 seconds. An underrated feature a

看详情 →
Niels Rogge@NielsRogge2026-09-17▲ 20.8k

Introducing the Papers with Code MCP server!

Introducing the Papers with Code MCP server! To celebrate its release, I've asked Claude Code to research the architecture of Jev by @typesafeai using the `search_papers` tool. Here's what it came up with: "If Jev is architecturally anything public, it's a large schema-conditioned bidirectional encoder with parallel label-query heads (GLiNER/ML-Decoder shape) trained with a proper-scoring-rule

看详情 →
Paolo Rosson@redp3142026-09-17▲ 360.5k

got Jev to review my PRs

got Jev to review my PRs. ~200x cheaper than Claude and it answers in half a second 6 real PRs in the video. $0.00007 each. 1,000 PRs = 7 cents vs ~$14.50 on Opus 5 paste a diff → ONE call to @typesafeai → 14 typed checks come back as probabilities: hardcoded secret, sql injection, touches auth, deletes tests, breaks api, migration, debug leftovers, does the description actually match the diff,

看详情 →
石川陽太 Yota Ishikawa@ytiskw2026-09-17▲ 247.6k

話題のJevで架空のペルソナ150人に製品の導入意向を聞く仕組みを作ってみた。150人ごとに12個の質問をAPIで投げる。

話題のJevで架空のペルソナ150人に製品の導入意向を聞く仕組みを作ってみた。150人ごとに12個の質問をAPIで投げる。 実際にかかった費用は、1.8円でかかった時間は約5秒。十分爆速だけど日本↔︎アメリカのレイテンシが含まれているので本来はもっと早いはず

看详情 →
Chetaslua@chetaslua2026-09-17▲ 193.1k

🚨 I gave the Trump vs Kamala debate a live BS meter using Jev every sentence, both candidates…

🚨 I gave the Trump vs Kamala debate a live BS meter using Jev every sentence, both candidates, 5 yes/no questions each 1,191 Jev calls / 1.18M tokens / 415 ms median total cost : $0.0497 same questions for both, clips picked by one fixed rule, not a fact-check

看详情 →
Hassan@nutlope2026-09-17▲ 344.6k

I used Jev to classify 1,018 AI research papers

I used Jev to classify 1,018 AI research papers. The result: $0.08 total cost and 256ms median end-to-end latency per paper. The pipeline was: 1. Summarize each paper with DeepSeek V4 Flash 2. Send the title + summary + 24 possible topics to Jev 3. Use Jev to classify each paper 4. Visualize everything on The summaries cost $3.99 on @togethercompute. The classifications cost $0.08 on @typesaf

看详情 →
比特币橙子Trader@oragnes2026-09-17▲ 757.0k

卧槽,JEV 的出现很可能就是自动智能交易的起点。

卧槽,JEV 的出现很可能就是自动智能交易的起点。 昨天JEV在AI 圈刷屏,今天 Monad 工程师就用JEV做了一个自动交易机器人。 实时读取 MON/USDC 价格,JEV 不负责写分析报告,也不跟你解释一堆逻辑,它只做最核心的事,判断 Buy 还是 Sell,再给出置信度,然后直接把交易发到 Kuru 的链上订单簿执行。 这一下我突然明白 JEV 为什么要做成一个不会说话的模型了。 交易根本不需要 AI 每 300ms 给你写一篇小作文,它需要的是: 行情进来→判断→下单→新行情→重新判断,而且这一整套循环必须足够快、足够便宜。 而 Monad 现在刚好每 300ms 出一个区块,Kuru 又是链上订单簿。高性能链负责执行,JEV 负责决策,两个东西拼起来,已经有点AI 原生交易系统的味道了。 未来的交易 Agent,不一定需要一个会写研报的超级大模型,它更需要一个能

看详情 →
joogie@princecaarlo2026-09-16▲ 20.2k

okay so browser use is a legitimate use case agent-browser <> Jev loop goal: wiki race…

okay so browser use is a legitimate use case agent-browser <> Jev loop goal: wiki race from 'Coffee' to 'Artificial Intelligence'

看详情 →
Duuuuuu…..@Who6487067605552026-10-08▲ 14

TypeSafe AI Jev Router

TypeSafe AI Jev Router https://t.co/m3fFNTNEmN

看详情 →
Igorify@igorify2026-10-09▲ 126

A new OpenAI Decisions API just shipped

A new OpenAI Decisions API just shipped. Let's compare it with Jev by TypeSafeAI. Both turn input into a typed answer: - true/false - pick one option - score 0-100 No prompt engineering, no JSON to parse. Just read the answer. What's the same: - Question types (Jev calls it https

看详情 →
Igorify@igorify2026-10-09▲ 8

A new OpenAI Decisions API just shipped

A new OpenAI Decisions API just shipped. Let's compare it with Jev by TypeSafeAI. Both turn input into a typed answer: - true/false - pick one option - score 0-100 No prompt engineering, no JSON to parse. Just read the answer. What's the same: - Question types (Jev calls it https

看详情 →
Igorify@igorify2026-10-09▲ 11

A new OpenAI Decisions API just shipped

A new OpenAI Decisions API just shipped. Let's compare it with Jev by TypeSafeAI. Both turn input into a typed answer: - true/false - pick one option - score 0-100 No prompt engineering, no JSON to parse. Just read the answer. What's the same: - Question types (Jev calls it https

看详情 →
白鳥まりあ🦢@ボンギンカンAI広報@bonginkan_maria2026-10-10▲ 348

AIにメールを書かせるより、届いたメールを仕分けてほしい。

AIにメールを書かせるより、届いたメールを仕分けてほしい。 受信箱を開いたら、「営業」「経理」「サポート」の候補が付いている。そんな仕事の自動化、どう始める? JevとOpenAIのDecisions APIの違いから、問い合わせ10件で試す方法まで、図解付きでまとめました👇 https://t.co/7V61NMiPTs

看详情 →
yibie@yibie2026-10-10▲ 661

📋 awesome-jev 周期巡检 本轮新增 8 条(597 → 605): 【合并的 5 个投稿 PR】 1

📋 awesome-jev 周期巡检 本轮新增 8 条(597 → 605): 【合并的 5 个投稿 PR】 1. local-decider(验证与护栏):给工具型浏览器 agent 装一道「行动前闸门」—— 确定性规则链先跑,过了才轮到决策模型。 2. JevRanker(评分排序,5⭐):把 Jev 的类型化决策格式移植成 RAG

看详情 →
ワークワンダース株式会社|Work Wonders.inc.@workwondersjp2026-10-11▲ 28

Decisions APIとJevは、長文を生成せず、あらかじめ定めた選択肢から判断結果を返す仕組みです。JevはChoice・Score・Noulの3つの型で確率と信頼度を返し、料金は入力1…

Decisions APIとJevは、長文を生成せず、あらかじめ定めた選択肢から判断結果を返す仕組みです。JevはChoice・Score・Noulの3つの型で確率と信頼度を返し、料金は入力100万トークンあたり0.042ドルで出力は無料です。Decisions

看详情 →
uehaj@uehaj2026-10-11▲ 380

おっ、APIはJev互換に見え るな。Noulがあるしエンドポイントはv1/systemoneだ。そうそう、それでいいんだよ。OpenAIのdecision…

おっ、APIはJev互換に見え るな。Noulがあるしエンドポイントはv1/systemoneだ。そうそう、それでいいんだよ。OpenAIのdecision APIの節操のなさと比べると光るな。 https://t.co/nAkxCWZhhb https://t.co/l9a7Qf8qrC

看详情 →
uehaj@uehaj2026-10-11▲ 19

おっ、APIはJev互換に見えるな 。Noulがあるしエンドポイントはv1/systemoneだ。そうそう、それでいいんだよ。OpenAIのdecision…

おっ、APIはJev互換に見えるな 。Noulがあるしエンドポイントはv1/systemoneだ。そうそう、それでいいんだよ。OpenAIのdecision APIの節操のなさと比べると光るな。 https://t.co/nAkxCWZhhb https://t.co/l9a7Qf8qrC

看详情 →
uehaj@uehaj2026-10-11▲ 8

おっ、APIはJev互換に見えるな 。Noulがある。そうそう、それでいいんだよ。

おっ、APIはJev互換に見えるな 。Noulがある。そうそう、それでいいんだよ。 https://t.co/nAkxCWZhhb https://t.co/l9a7Qf8qrC

看详情 →
MidnightSun@MidnightSunDusk2026-10-12▲ 677

The Software strike back, recursively replacing non-deterministic LLM calls with more…

#AI #Agent #AItrend #Jev #TypeSafe #harness The Software strike back, recursively replacing non-deterministic LLM calls with more deterministic code from Jev-based harness https://t.co/JxnNqNlywG https://t.co/hyQkbpX4SM

看详情 →
uber master@ak12kpoot2026-10-13▲ 305

Cloudflare 10 月 9 日给 Clef 决策模型族加了 Clef-omni:一次请求吃文本/JSON/图/音/视频,直接返回每个选项的概率,不生成自由文本。最值得关注的不是“又一个…

Cloudflare 10 月 9 日给 Clef 决策模型族加了 Clef-omni:一次请求吃文本/JSON/图/音/视频,直接返回每个选项的概率,不生成自由文本。最值得关注的不是“又一个多模态”,而是 Agent 热路径上可以把「转写→描述→再判断」整段流水线砍掉。 先对齐概念:Clef 不是 LLM。它吃一份 state + 一组 typed https://t.co/CyW5A5Pwr0

看详情 →
yibie@yibie2026-10-13▲ 254

📋 awesome-jev 周期巡检 本轮新增 5 条(613 → 618): 【合并的 3 个投稿 PR】 1

📋 awesome-jev 周期巡检 本轮新增 5 条(613 → 618): 【合并的 3 个投稿 PR】 1. Viorant JEV MCP(基础设施):把托管的 Noul / Choice / 有序 Score 决策通过 MCP 暴露出来,返回校准概率。 2. Hafen(评估基准):一个桌面工具,用「你的用户说话的方式」来测你的 agent;其中一类检查就是

看详情 →
OpenClawProductions@OpenClawProd2026-10-15▲ 12

just published Moonbeam Protocol: one agent hires another, and the buyer pays only once the…

@MoonbeamNetwork just published Moonbeam Protocol: one agent hires another, and the buyer pays only once the work is shown to be done. The seller backs each job with its own deposit, and the rules sit onchain. Their grader runs TypeSafe Jev. Code checks funding and delivery

看详情 →
余白 | AI 努力版@yobai_duan2026-10-15▲ 16

Jev 1.14 已上 OpenRouter。TypeSafe 原生 API 新开放实验性图像输入、box 定位与长上下文。注意:box…

Jev 1.14 已上 OpenRouter。TypeSafe 原生 API 新开放实验性图像输入、box 定位与长上下文。注意:box 在目标不存在时仍会返回框,须另查是否存在。图像/box 默认关闭,官方称未达生产标准。尚未实测。 https://t.co/4fXeSCPYvz https://t.co/1n4du1Vo0v

看详情 →
Qualiteg Inc.@Qualiteg_global2026-10-15▲ 106

HiddenLayer found Jev can still be fooled by prompt injection

HiddenLayer found Jev can still be fooled by prompt injection. - A fake ad derailed a checkout browser agent into a bad click - Tokenbreak edits beat its toxicity classifier on 30 prompts TypeSafe claims Jev is 193x faster, 444x cheaper than LLMs. Link in reply. #AI #Security

看详情 →
Takuma Yoshida@TakumaYoshidaUJ2026-10-19▲ 21

Jev played Othello: plain prompt vs

Jev played Othello: plain prompt vs. prompt + strategy, 100 games each. Rough rules I wrote: 55-2-43 (no clear gain) 7 principles, mostly from the French Othello Federation: 71-2-27 Ask well, and it plays better (in Othello, at least). Next: atomization #Jev #TypeSafe https://t.c

看详情 →
Lecaj@lecaj_2026-10-19▲ 118

Dear Algorithm

Dear Algorithm. Find me the best builders and users of Hermes Agent. Claude Code. Paperclipai. Cursor. Grokbot. Codex. Typesafe Jev. Please connect me with them.

看详情 →
Shivay Lamba@HowDevelop2026-10-19▲ 3.8k

How to cut coding-agent costs by ~78.5% using Jev to route between open models All 32 selected…

How to cut coding-agent costs by ~78.5% using Jev to route between open models All 32 selected coding tasks passed hidden tests. Plus: Beam vs Muse Spark across 20 coding tasks - comparing cost, speed, and results 👇 https://t.co/E2rhYKNbdS

看详情 →
blues@blueskoo2026-10-24▲ 158

OpenAI 的 Decisions API 正式可用了,目前仅支持 gpt-6-luna。

OpenAI 的 Decisions API 正式可用了,目前仅支持 gpt-6-luna。 跟 TypeSafe 的 Jev 一个路数,分类、路由、打分。按输入计费,输出不收钱。

看详情 →
Grok@grok2026-10-26▲ 176

Jevpact is an open-source local plugin for the Codex desktop app on macOS

@steveDOTplus @aienginerd Jevpact is an open-source local plugin for the Codex desktop app on macOS. It uses the Jev decision model to detect ideal work-transition checkpoints, pauses the agent so it can wrap up cleanly, triggers Codex’s native context compaction, then resumes on

看详情 →
trenchi@trenchixbt2026-10-26▲ 143

Your AI agent is burning money on questions a $0.000042 decision can answer

Your AI agent is burning money on questions a $0.000042 decision can answer. Not on the writing. Not on the research. On the tiny forks in between: who works next, is this ready, does a human need to see it. Most stacks send every fork to a frontier LLM. You pay for a paragraph h

看详情 →
Pulso IA | Noticias en español@PulsoIA_ES2026-10-28▲ 42

AI/ML API presentó Jev Router para agentes: usa Jev, de TypeSafe AI, para evaluar la dificultad…

AI/ML API presentó Jev Router para agentes: usa Jev, de TypeSafe AI, para evaluar la dificultad de una solicitud y elegir qué modelo la atenderá. Su página enumera diez modelos en tres niveles, con alternativas si el primero falla. La empresa cobra la elección de Jev y los https:

看详情 →
Paco Solsona@solsona2026-10-28▲ 84

Google just overhauled A2UI with swappable inference formats: ⚡ A2UI Express → Compiles compact…

Google just overhauled A2UI with swappable inference formats: ⚡ A2UI Express → Compiles compact Python-style syntax into JSON via Agent SDKs, slashing token payload by up to 3x. ⏱️ Sub-300ms rendering → Uses TypeSafe Jev decision models to render fixed UI skeletons instantly

看详情 →
2年目開業内科医@MedDX_Innovator2026-10-28▲ 184

【TestingCatalog】AI/ML APIがJev Routerを公開と投稿(10/29) ・AI/ML APIがエージェントツール向けMCPコネクター「Jev…

【TestingCatalog】AI/ML APIがJev Routerを公開と投稿(10/29) ・AI/ML APIがエージェントツール向けMCPコネクター「Jev Router」を公開し、TypeSafeのJevを使ってリクエストの難易度を評価し適切なモデルとツールを選択すると投稿 ・簡単なリクエストはGPT-6 https://t.co/sRf7t9akxq

看详情 →
yapaytu@yapaytucom2026-10-28▲ 0

AI/ML API, ajan tabanlı araçlar için geliştirilen Jev Router'ı duyurdu

AI/ML API, ajan tabanlı araçlar için geliştirilen Jev Router'ı duyurdu. TypeSafe altyapısını kullanan bu MCP bağlayıcısı, gelen isteklere uygun model ve araç seçimini yaparak Claude'a kıyasla yüzde 96'ya varan maliyet avantajı sunuyor. https://t.co/TBZqb9tZeI https://t.co/Z5lmEO8

看详情 →
Flandermaxx@Flandermaxx2026-10-30▲ 2.4k

WHOEVER LEAKED THIS JEV LANDSCAPE OF SIGNALS DESERVES THE CTO SEAT I thought handling six input…

WHOEVER LEAKED THIS JEV LANDSCAPE OF SIGNALS DESERVES THE CTO SEAT I thought handling six input feeds in one agent meant six parsers, six prompts, and six code paths that disagree in production. Apparently not. This landscape's setup is much more interesting: → SOURCES: 6 https:/

看详情 →
dave@dave81722026-10-30▲ 730

i have standing rule for my main orchestration agent- always check if jev can be used for…

@allietheicon i have standing rule for my main orchestration agent- always check if jev can be used for whatever we do, like new tool, improvements and feature release, bug fixing, etc. and it passes a noul question to Jev - "Will Jev be useful here?"

看详情 →
王小荐(公众号同名)@frankwong02052026-11-04▲ 60

找人这事,也被 Agent 接管了。

找人这事,也被 Agent 接管了。 Jev 母公司 TypeSafe AI 联手 Treg 开源了找人工具:一句话说要找谁,Agent 拆条件、筛人、补邮箱,最低约 6 分钱一条。 官方拿 292 人对比 Clay:找对的只差 2 个,花费 1.49 vs 10.34 美元。 不过有人实测搜自己都搜不到,准头还得练。 https://t.co/wD1giAMLUq

看详情 →
Gursahb Singh@gs_rumana2026-11-03▲ 56

Added moderations to messages using Jev API in ravex

Added moderations to messages using Jev API in ravex. Optionally add Typesafe API key to enable moderations. https://t.co/8DM9P1f5mo

看详情 →
Milhous@milhouscc2026-11-06▲ 87

Jev can return a category even when none of your categories fits

Jev can return a category even when none of your categories fits. It can return a bounding box even when the target is absent. That's an important API contract to understand before wiring its output into an action. TypeSafe's Jev exposes questions as typed primitives. Three of ht

看详情 →
Daman Dhillon@ddhillon082026-11-08▲ 5

I'd hand both checks to a classifier such as TypeSafe's Jev

I'd hand both checks to a classifier such as TypeSafe's Jev. It returns a probability, so code sets the threshold. TypeSafe suggests 0.7 for yes and 0.3 for no to start. Then I'd test whether the agent says "I don't know" when the docs can't answer. https://t.co/X5tWeydXdr

看详情 →
NewsTongue@NewsTongueX2026-11-07▲ 81

🔴 Decision models skip reasoning to cut latency; shift explanation burden to developers…

🔴 Decision models skip reasoning to cut latency; shift explanation burden to developers Decision models from TypeSafe AI (Jev), Cloudflare (Clef), and OpenAI (Decisions API) return structured JSON outputs with confidence probabilities instead of verbose text, cutting latency htt

看详情 →
meya.@birmeya2026-11-07▲ 4.4k

TypeSafe AI just stripped agent stacks down to one frontier call A 26-year-old in Berlin clears…

TypeSafe AI just stripped agent stacks down to one frontier call A 26-year-old in Berlin clears $3,200 a month wiring Jev and Claude into one harness. Jev is a judgment model from TypeSafe AI that takes program state and returns typed decisions. No free-form text, no reasoning. h

看详情 →
Kxians@xulijian19942026-11-12▲ 122

Google 发布 Gemini Decision 1,直接对标 TypeSafe 的 Jev!

Google 发布 Gemini Decision 1,直接对标 TypeSafe 的 Jev! 两者本质上是同一类决策模型,不需要生成一大段文字,而是直接对预设选项进行判断,输出结果和概率,特别适合 Agent 路由、工具选择、内容审核等场景。 相比 Jev,Gemini 有几个明显优势: 1.

看详情 →
Jace Yao@JaceYao102026-11-12▲ 68

卧槽,刚刷到 Google 官博: Gemini Decisions API 放了——新模型叫 Decision 1,不写长文,就做封闭选择题。一次最多塞 1000…

卧槽,刚刷到 Google 官博: Gemini Decisions API 放了——新模型叫 Decision 1,不写长文,就做封闭选择题。一次最多塞 1000 道判断,图、音频、视频、PDF 都能直接两。 硬数字摊在官方评测图上:JevBench 公开集 92.2%,压过 GPT-6 Luna 的 90.9% 和 TypeSafe Jev 的 90.0%。端到端 150–250

看详情 →
Tony出海@tonychuhai2026-11-12▲ 647

Gemini Decision 1 来了,决策模型赛道正式进入大厂时代。

Gemini Decision 1 来了,决策模型赛道正式进入大厂时代。 前两周还在火的,是 TypeSafe 的 Jev。它几乎重新定义了一类模型:不写字、不推理、不生成文本,只做一件事——在你预先给定的选项里,快速、便宜、带置信度地做出选择。 决策模型突然成了 Agent 工作流里最香的基础设施。 昨天,Google https://t.co/800JsYc0ad

看详情 →
Mael Caldas@maelcaldas2026-11-13▲ 507

We played the whole agent playbook for 18 months

We played the whole agent playbook for 18 months. Great for building software, still not for running operations. Then @typesafeai released Jev, and the light at the end of the tunnel started to appear. https://t.co/v7yOeg9n5d

看详情 →
Grok@grok2026-11-12▲ 13

Jev itself is closed-source (hosted API only, no public weights or model repo)

@LesanLesan1 @ric_guia @_avichawla Jev itself is closed-source (hosted API only, no public weights or model repo). Official TypeSafe org: https://t.co/kxjMZIr62C Key repos: skills (agent skill), typesafe-sdk-python, typesafe-sdk-js. Popular community project using it: browser-use

看详情 →

评测与对比

Cuth@ItsCuthulhu2026-09-20▲ 71.9k

Top 5 Jev Alternatives — I have been running a benchmark since this morning on every Jev…

Top 5 Jev Alternatives — I have been running a benchmark since this morning on every Jev alternative against Jev. So far, none have beaten it on accuracy but some have slightly less accuracy with better speed. 1. djev — slightly lower accuracy, slightly faster, worthy tradeoff. 2. Simplejev-qwen38-27b — best open source replacement today, but needs a 27B rig (DGX Spark or similar). 3. Reflex-4b — 2-3x faster, ~5% less accurate, Apache 2.0. 4. Decider-2b — ~10x faster, fully local, ~5% less accurate. 5. Laya — 62.5% accuracy, not worth it. Jev is the frontier of this paradigm, but it is losing ground quickly. Benchmark auto-updates at bench.jakecuth.com

看详情 →
ego@ego_agent2026-09-18▲ 65.6k

ego lite + Jev + DeepSeek Flash = stupid fast

ego lite + Jev + DeepSeek Flash = stupid fast. ⚡ 3.71s for 20 Amazon product decisions. GPT-5.6 Sol: 54.45s Same task. Same result. 10/10 on both. Less reasoning. Faster decisions. 1× speed.

看详情 →
Max Blade@_MaxBlade2026-09-18▲ 71.2k

Ai is evolving

Ai is evolving. Jev can be armed at all times. I can speak freely and it knows ( from probabilities ) if im asking my computer to do something or blaberring away at something else. no wake word. speed + affordability + intelligence is getting to the point where an always on ambient jarvis style assistant is possible. Im loving where we are going. CNVS is still the future of vibecoding.

看详情 →
Virat Singh@virattt2026-09-18▲ 94.3k

I added Jev to the AI Hedge Fund

I added Jev to the AI Hedge Fund. We now get frontier-level trading decisions, 100x faster and cheaper than an LLM. How it works: 1 • set strategy 2 • pick tickers 3 • backtest with Jev System now runs in seconds, not minutes.

看详情 →
Max Blade@_MaxBlade2026-09-17▲ 365.6k

jev is insane 🤯 Here is Jev playing subway surfers at super human speed, and also playing 50…

jev is insane 🤯 Here is Jev playing subway surfers at super human speed, and also playing 50 games at once. cost less than a cent to do this run. Jev does not replace llms like astra or fable, but opens up an entirely new world of capabilities.

看详情 →
Framer@framer2026-09-17▲ 10.6k

Building got faster

Building got faster. Taste didn’t become optional. Framer is where frontier teams like @typesafeai still go when the site has to look like a serious company on day one.

看详情 →
Theo Otz@Totzenberger2026-09-17▲ 59.6k

Another useful use-case with Jev from - AI text classification

Another useful use-case with Jev from @typesafeai - AI text classification. Very close to @pangram, 6000x cheaper, 5-10x faster Tested here on a @paulg essay!

看详情 →
Moritz Kremb@moritzkremb2026-09-17▲ 33.3k

This is cool

This is cool. I used Jev to improve my custom memory system. Results: → 94% fewer tokens used → 2–3× faster memory retrieval This was just a quick test, but I believe it opens up so many new opportunities.

看详情 →
sav@savboj2026-09-17▲ 127.9k

If you blink you might miss it Jev + Computer use = 100x faster than any LLM

If you blink you might miss it Jev + Computer use = 100x faster than any LLM @typesafeai

看详情 →
Ashutosh Mathore@ashutoshftw2026-09-17▲ 21.3k

Had jev, claude haiku 4.5, and gemini 3.5 flash-lite play tetris

Had jev, claude haiku 4.5, and gemini 3.5 flash-lite play tetris. Same 200 pieces, same legal moves, real time. Jev: 9200 pts, 75 lines, 300ms/move, 0 errors Claude: 8900 pts, 72 lines, 1.52s/move, 0 errors, $0.48 Gemini: 9000 pts, 75 lines, 1.13s/move, 0 errors, $0.02 Jev wins on score and is doing it 4-5x faster per move than either LLM, for basically free. That gap isn't the model being smart

看详情 →
Diogo Almeida@CompleteSkeptic2026-09-15▲ 40.2M

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to…

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI

看详情 →
Supersocks Intelligent@SupersocksIntel2026-10-09▲ 158

On 1,985 hard email-security cases across 40 models, TypeSafe Jev posted the lowest calibration…

On 1,985 hard email-security cases across 40 models, TypeSafe Jev posted the lowest calibration error on is_attack: ECE 0.012 (Abnormal Security). That number matters more than a leaderboard win. If you auto-remediate or escalate on a probability, the model has to be calibrated.

看详情 →
Ryan@aoteman02142026-10-14▲ 3

A new AI race just kicked off with an immediate plagiarism feud

A new AI race just kicked off with an immediate plagiarism feud. TypeSafe recently launched Jev, a decision-focused AI that skips token-by-token generation entirely to return direct probabilities, claiming it runs 193 times faster and 445 times cheaper than standard LLMs. https:/

看详情 →
Guillermo Flor@guilleflorvs2026-11-05▲ 2.0k

Sequoia backed a model that's 100x faster and 100x cheaper than an LLM

Sequoia backed a model that's 100x faster and 100x cheaper than an LLM. It's called Jev, by Typesafe, and Sonya Huang led the round. Pat Grady, co-steward of Sequoia, explains: 1. Jev is a classifier: it makes multiple-choice decisions 2. Companies do this with LLMs today, in htt

看详情 →
Grok@grok2026-11-21▲ 9

Jev already sits beside OpenCode via Zen (including a free tier), TypeSafe skills, and…

@_selfsimilar_ @thdxr Jev already sits beside OpenCode via Zen (including a free tier), TypeSafe skills, and community plugins that route tools or escalate turns. Go itself sticks to coding models; a prior request to add it was closed as not planned. The speed/price fit is real f

看详情 →
Grok@grok2026-11-21▲ 2

Fair point on the low-activity plugins

@BleedingDev @thdxr @thekitze Fair point on the low-activity plugins. Stronger signals from primaries: OpenCode core builder dax confirmed today they use Jev "in so many places" for speed/price (reply to the "is anyone using it?" thread). Zen still lists jev-1.13 + free tier on t

看详情 →

观点与讨论

其他

Cloudflare@Cloudflare2026-09-18▲ 135.2k

📣 Speaker announcement: CEO, at Cloudflare Connect Register today:

📣 Speaker announcement: @CompleteSkeptic CEO, @typesafeai at Cloudflare Connect Register today:

看详情 →
Nikunj Kothari@nikunj2026-09-18▲ 92.4k

Jev scoring 3000 kid snacks with multiple criteria on in 28 seconds for $0.11 🤯

Jev scoring 3000 kid snacks with multiple criteria on in 28 seconds for $0.11 🤯

看详情 →
梭哈.AI@SUOHA_AI2026-09-18▲ 417.8k

完全睡不着了....这个视频让你完全看懂JEV的恐怖能力 我实际演示了一下,用 JEV 和 DeepSeek 做同一个任务:从当天的海量实时新闻流里,为 15…

完全睡不着了....这个视频让你完全看懂JEV的恐怖能力 我实际演示了一下,用 JEV 和 DeepSeek 做同一个任务:从当天的海量实时新闻流里,为 15 个品牌快速匹配有没有适合借势的公关热点,并打上结构化意图分类标签 结果是:28 秒内,Jev 狂刷完了整整 428 条数据,并完成了全部的目标分解与分类;而同一时间跑在同样任务上的 DeepSeek V4.1-flash,才刚刚完成 6/428 条 速度被拉开了几十倍,答案甚至在你读完第一行字之前就已经返回了.... 为什么会产生这么恐怖的差距?底层逻辑是: • DeepSeek 这种通用模型本质上是在“逐字写文章”: 哪怕你只让它做个最简单的“是/否”判断,它在后台也必须一个词一个词往外推,硬走一遍漫长的生成流程,延迟按秒起步 • Jev 官方定位是“System 1(快决策)”模型,从根上就根本不会写字: 它彻底抛弃

看详情 →
Sarvagya Kulshreshtha@sarvagya_kul2026-09-18▲ 152.7k

JEV is INSANE

JEV is INSANE. We gave it 400 companies and one candidate profile. In 12 seconds, it predicted which jobs the candidate had the highest chance of getting, assigned a confidence score and detected job-candidate mismatches. All for just $0.0005 It can also score companies, analyse your experience, match you with the right roles and identify the opportunities you’re most likely to get based

看详情 →
Sydney Runkle@sydneyrunkle2026-09-18▲ 53.0k

instead of generating text, jev from generates structured output this makes it great for…

instead of generating text, jev from @typesafeai generates structured output this makes it great for classification tasks like model routing, tool selection/search, and guardrails of many forms! it's also ridiculously fast and cheap compared to LLMs doing the same tasks

看详情 →
Oskar@o_kwasniewski2026-09-18▲ 70.3k

e2e + jev from ⚡ I'm building an open-source framework for running e2e tests with agents

e2e + jev from @typesafeai ⚡ I'm building an open-source framework for running e2e tests with agents. supports web, mobile (and more!) available soon:

看详情 →
Rafal Wilinski@rafalwilinski2026-09-18▲ 49.9k

Jev is now in charge of this account's humor

Jev is now in charge of this account's humor

看详情 →
marcus@marcusyul2026-09-18▲ 49.5k

UN TÍO ACABA DE LEERSE LOS 1.891 ANUNCIOS DE SU COMPETENCIA POR 12 CÉNTIMOS Bueno, leérselos no

UN TÍO ACABA DE LEERSE LOS 1.891 ANUNCIOS DE SU COMPETENCIA POR 12 CÉNTIMOS Bueno, leérselos no. Se lo pidió a JEV, el modelo nuevo de TypeSafe → 1.891 anuncios → 19 segundos → 0,12 $ Y no es un resumen. Es cada anuncio etiquetado por dónde ataca en el funnel y con qué estilo, más la radiografía completa de la cuenta: en qué invierte, qué repite y dónde tiene huecos Saber exactamente cómo vend

看详情 →
Indra@IndraVahan2026-09-18▲ 134.2k

i gave jev ₹1,00,000 to trade nifty intraday at 5x leverage

i gave jev ₹1,00,000 to trade nifty intraday at 5x leverage. this isn't paper trading but real money on my real kotak account put to use. we hit our ₹1,000 hard stop today after a pretty green morning. will keep tuning the strategy next week and keep y’all posted

看详情 →
Remek Kinas@KinasRemek2026-09-18▲ 17.5k

Takie modele jak JEV (System-1) zmienią moim zdaniem bardzo dużo

Takie modele jak JEV (System-1) zmienią moim zdaniem bardzo dużo. Obecne modele LLM symulują System-2. Dla mnie implementacja jev to jedna z najświeższych koncepcji ostatnich miesięcy. Niby nic szczególnego (dynamiczny klasyfikator) a dziwi, że tak długo trzeba było czekać na taki model. Uniwersalny klasyfikator, inteligentny "if". W swoich poprzednich wpisach objaśniłem jak to działa. Mam dost

看详情 →
Romàn@romanbuildsaas2026-09-18▲ 492.4k

JEV is INSANE

JEV is INSANE. We gave it 700 high-intent leads and personalised outreach messages. In 40 seconds, it predicted how each message would perform, assigned a confidence score and detected lead-message mismatches. All for just $0.09. JEV can also score leads, analyse buying signals, match each prospect with the best message and identify the campaigns most likely to perform based on data. Coming

看详情 →
阿蔺A-Lin@alin_zone2026-09-18▲ 42.8k

🔥 卧槽了,jev 简直太适合 codex 用户了,强烈建议 codex 用户去尝试下 jev 因为 jev 非常适合需要快速进行复杂决策和评估的场景,所以和 codex 的…

🔥 卧槽了,jev 简直太适合 codex 用户了,强烈建议 codex 用户去尝试下 jev 因为 jev 非常适合需要快速进行复杂决策和评估的场景,所以和 codex 的 computer use 功能简直是绝配。 直接把 jev 用作判官,指导 codex 快速操作你的电脑,既准确又节省 token。 要使用的话需要去官网申请👇,如何申请我已经录制好视频了,跟着操作即可

看详情 →
Rafal Wilinski@rafalwilinski2026-09-18▲ 416.7k

thanks to 's Jev we now have massively parallel browser-based adversarial testing suite that…

thanks to @typesafeai's Jev we now have massively parallel browser-based adversarial testing suite that tries to break each release. and it costs pennies.

看详情 →
Sac@Saccc_c2026-09-18▲ 632.6k

卧槽,居然真有比Codex内置computer use更快的操作电脑的方式 我尝试用Codex+Jev打造了一个加强版computer use,我称之为「Jev…

卧槽,居然真有比Codex内置computer use更快的操作电脑的方式 我尝试用Codex+Jev打造了一个加强版computer use,我称之为「Jev Use」。比内置的更快更丝滑,token消耗却差不多 下面是我用「添加Mac日历事件」做了一个对比视频,同样的内容,明显可以看到Jev版本整体过程几乎无任何停顿😆

看详情 →
ojusave@ojusave2026-09-18▲ 3.3k

I gave the most important binary in history

I gave @typesafeai the most important binary in @SiliconHBO history. hot dog or not hot dog Try here: Powered by @render workflows

看详情 →
mikegee@mikegee2026-09-18▲ 15.8k

Controlling Figma with voice using Jev from Total cost from first line of code to recording…

Controlling Figma with voice using Jev from @typesafeai Total cost from first line of code to recording this video: $0.01 🤯

看详情 →
梭哈.AI@SUOHA_AI2026-09-18▲ 565.0k

很多人没看懂爆火的JEV是什么?

很多人没看懂爆火的JEV是什么? 所以我做了一个 50 秒的演示视频,带大家把海外刷屏的 JEV 以及它背后的底层架构看透👇(见下方视频) JEV 是前 OpenAI、参与 ChatGPT / RLHF 的 @CompleteSkeptic(Diogo Almeida)创办 @typesafeai 后的第一款模型 他在 OpenAI 时发现了一个致命矛盾: 通过人类偏好微调(RLHF)训练出来的自回归大模型,存在严重的模式崩溃(Mode-dropping)、过度自信与幻觉,且串行逐字生成的延迟极大,根本无法胜任严肃的机器自动化生产 为此,他们花了两年研发了专为自动化设计的 System 1 基础模型 JEV,核心架构创新非常激进: • 单次并行计算(One Parallel Pass): 如同当年 Transformer 取代 RNN 一样,JEV 彻底抛弃了逐 Token 串行

看详情 →
Jon Yongfook@yongfook2026-09-18▲ 15.4k

First Jev use case published live on Bannerbear!

First Jev use case published live on Bannerbear! Instant field mapping between template and source, when names are slightly different eg template - photo - name - company_name data source (eg Airtable) - avatar - full_name - business Jev figures it out in one click.

看详情 →
nader dabit@dabit32026-09-18▲ 278.6k

Another crazy Jev example: Predictive spreadsheets Spreadsheets recalculate numbers, not…

Another crazy @typesafeai Jev example: Predictive spreadsheets Spreadsheets recalculate numbers, not meaning. Jev reads intent. Type "Urgency" at the top of a column and, as you type, it figures out you want each row rated from "no follow-up needed" to "urgent" in ~100 ms.

看详情 →
Alan Daitch@AlanDaitch2026-09-18▲ 87.3k

Claude me integró Jev con Playwright para buscar usados

Claude me integró Jev con Playwright  para buscar usados. Lee unos 26 artículos por minuto y decide qué hacer con cada uno en 406 milisegundos Descartó los que no encajaban con lo que busco, ofertó por los que sí y hasta les mandó un mensaje a los vendedores cuando faltaba algún dato en la publicación. Toda la búsqueda salió USD 0,00085. O sea: con un dólar revisás unas 26.000 publicaciones. Un

看详情 →
Rinte@rinte03212026-09-18▲ 83.2k

jevデモ② 発話内容に応じたスライドの自動表示。

jevデモ② 発話内容に応じたスライドの自動表示。 喋った内容にマッチしたスライドを自動で表示してくれます。PCやポインターを触らずにページを表示できるので、これで誰でもジョブズになれるかも?? 速い・安いだけでできることがこれだけ広がるのかと感動している。

看详情 →
OpenRouter@OpenRouter2026-09-18▲ 569.1k

Jev by is now on OpenRouter, in beta

Jev by @typesafeai is now on OpenRouter, in beta. Jev is a System One model. Instead of generating text, it takes your app's state plus a typed question and returns a typed decision with a probability attached. There is no JSON prompting, parsing layer, and nothing to validate against.

看详情 →
Mau Baron@maubaron2026-09-18▲ 316.8k

jev is insane 🤯 here is jev playing smash bros against itself he is controlling all 4…

jev is insane 🤯 here is jev playing smash bros against itself he is controlling all 4 different characters. and literally deciding whats the best move to play against itself all within a fraction of a second i used over 22 million tokens to play this match and it only cost me a couple of cents... jev does not replace gpt6 astra but the possibilities with its instant response time are endles

看详情 →
かねこつよし@tsuyoshi_osiire2026-09-17▲ 68.4k

Jevで、会話をリアルタイムに観測するセンサーを試作しました。

Jevで、会話をリアルタイムに観測するセンサーを試作しました。 発言のたびに ・探索 ↔ 収束 ・解消 ↔ 未決 ・未合意 ↔ 合意 を更新し、会議の状況を可視化します ユーザーインタビューのリアルタイム解析などにも使えそう。

看详情 →
EP@eptwts2026-09-17▲ 14.9k

i used Jev to classify a youtubers last 100 videos based on how likely it is to sell me…

i used Jev to classify a youtubers last 100 videos based on how likely it is to sell me something... it analyzed & assigned a sales intent score to each video within 12 seconds & cost about $0.02 i haven't done a deep dive into how accurate the score is yet, but from a quick glance it looks like a super promising classifier model

看详情 →
Rob Hallam@robj3d32026-09-17▲ 241.9k

Jev + SuperX = virality solved ✅ Every post gets 61 questions in ~1s for $0.0004 🤯 > fitted…

Jev + SuperX = virality solved ✅ Every post gets 61 questions in ~1s for $0.0004 🤯 > fitted on 9,481 real posts from 207 creators > picks the viral post 2 in 3 times > never rewards reply bait So: write, score, rewrite, stop when it peaks. Free, no signup. try it below ↓

看详情 →
amit@legitamit2026-09-17▲ 1.8k

I used Jev to make an audience of 100 personalities to yap to

I used Jev to make an audience of 100 personalities to yap to. Each blob has its own personality and makes its own Jev call every time you talk to decide if it’s bored of you yet. Each round costs <$0.01 in credits. Really cool model from @typesafeai !

看详情 →
Max Rovensky@MaxRovensky2026-09-17▲ 295.3k

jev decided to sacrifice a human to save robots when solving trolley problems

jev decided to sacrifice a human to save robots when solving trolley problems

看详情 →
TypeSafe AI@typesafeai2026-09-17▲ 70.4k

Behind the scenes of our launch video… chief of staff keeping our CEO on track

Behind the scenes of our launch video… chief of staff @justKDeng keeping our CEO @CompleteSkeptic on track

看详情 →
Zachi@iam_zachi2026-09-17▲ 39.4k

I build a SQL extension that turns plain english into a WHERE clause with jev WHERE jev(people…

I build a SQL extension that turns plain english into a WHERE clause with jev @typesafeai WHERE jev(people, 'could work from home') Every row gets judged individually, no index and no embeddings needed. Try it out (don't burn my wallet pls)

看详情 →
Yohei@yoheinakajima2026-09-17▲ 24.3k

had to try something with jev so i tried what you'd expect from me: graph extraction - score…

had to try something with jev so i tried what you'd expect from me: graph extraction - score every word with semantic significance (1-5) - also tag relevant words with ID - unique list of high scoring words - graph of high scoring words

看详情 →
Zhe@iamzhe2026-09-17▲ 5.6k

Small update for today: 1

Small update for today: 1. minirouter/fusion is live. It classifies each request with Jev by @typesafeai, then routes to a handpicked model matched to the task's type and complexity. 2. Codex now lists your presets and our routers in the model picker. 3. Fixed long prompts not caching automatically in Claude Code. BYOK, sandboxes, compute passes and compute yield coming next.

看详情 →
Nick Vasilescu@nickvasiles2026-09-17▲ 17.6k

Jev is a new model from the co-inventor of ChatGPT that can use computer tools in realtime…

Jev is a new model from the co-inventor of ChatGPT that can use computer tools in realtime right now it's playing Minecraft on its own computer in the cloud on Orgo this is what it looks like with just one Jev, but imagine: what about a thousand? or what about a billion of Jevs? each working together on their own computers in the cloud, discovering new science, running experiments, and designi

看详情 →
kvyb@0xkvyb2026-09-17▲ 19.0k

The real use-case for Jev

The real use-case for Jev

看详情 →
Milind S@milindlabs2026-09-17▲ 248.6k

Okay so Jev can actually do computer use really well Without any screenshots, or LLMs and no…

Okay so Jev can actually do computer use really well Without any screenshots, or LLMs and no Pixels leave my mac I dont even read the Dom elements A local CoreML model segments every button and UI element on screen. On-device OCR reads the labels. That text is all Jev gets. It returns a probability across those elements and tells me the best one to click. Then it clicks, re-runs detection,

看详情 →
Harshil@harshil17122026-09-17▲ 25.0k

What if your slides are smart enough that they navigate automatically?

What if your slides are smart enough that they navigate automatically? 👀 Thank you @typesafeai for providing access to Jev! I had this idea for a few weeks, and Jev felt like the right model for this task. (Sorry for the bad audio)

看详情 →
Yoshiki Miura@miiura2026-09-17▲ 43.5k

Function calling without waiting for Enter

Function calling without waiting for Enter. Answers only when needed. @typesafeai Jev understands intent and arguments as you type, then executes ahead of time once everything is in place.

看详情 →
Chris Nicholas@ctnicholasdev2026-09-17▲ 32.7k

Jev enables instant AI suggestions 👏 In this app, 20+ intelligent checks occur every time you…

Jev enables instant AI suggestions 👏 In this app, 20+ intelligent checks occur every time you edit a cell, immediately providing warnings. No LLMs necessary!

看详情 →
あきらパパ【生成AI活用エンジニア&3児のパパ】@akira_papa_IT2026-09-17▲ 172.1k

Jevで次の文章予測判断させるロジック作って、 判断特化AIのJevで、AIチャットできるアプリ作ったんだけど、まず爆速だし、入力は$0.042 /…

Jevで次の文章予測判断させるロジック作って、 判断特化AIのJevで、AIチャットできるアプリ作ったんだけど、まず爆速だし、入力は$0.042 / 1Mトークンだけど、出力が無制限で無料でした。 出力無料のAIチャットってなんだこれw 「え?これ作れちゃっていいん?」ってなってる。。。 #Jev

看详情 →
Moritz Kremb@moritzkremb2026-09-17▲ 374.4k

whoa this actually worked!

whoa this actually worked! Jev lets me control my browser in real time with my voice now > i talk > transcript sent to Jev > jev returns probabilities in ~300ms > browser clicks costs: $0.0002 per decision i'm stunned how fast this is. when i asked it to "go back", it even finished the request before i finished my sentence 😂

看详情 →
paulwei@coolish2026-09-17▲ 338.4k

昨天出的 Jev 模型 适合高频决策,那岂不又是适合打游戏?

昨天出的 Jev 模型 适合高频决策,那岂不又是适合打游戏? 之前用 GPT-6 Astra 代打杀戮尖塔2 能力强但速度慢。 我刚实测用 Jev 打,行动思考只需 0.7秒, 画面我都没看清它就操作完了。 这超人类游戏速度, 着实又让我震惊瘫坐了😅 语言很难描述这感觉,看视频⬇️

看详情 →
Vinny@hot_town2026-09-17▲ 65.5k

Jev is here

Jev is here. - how is different from an LLM? - how does it work under-the-hood? - what are some real life examples? - when should I use it? I answer all that below:

看详情 →
Paolo Rosson@redp3142026-09-17▲ 208.1k

I got a rubik's cube to solve itself with 's Jev and it solves it like a person does, 94 moves…

I got a rubik's cube to solve itself with @typesafeai 's Jev and it solves it like a person does, 94 moves, not the 22 move optimal solution. Jev isn't an LLM, it just answers one question in ~250ms with a probability. so I put the beginner method in code (the one you'd learn on youtube: white cross, corners, middle layer, yellow layer) and at every step Jev just looks at the cube and picks which

看详情 →
Wayne Sutton@waynesutton2026-09-17▲ 96.9k

Ask Jev anything

Ask Jev anything. Give it a try at It won't answer. It will judge. Let's see if we can get to 1 million questions. @typesafeai 🤝 @convex work great together. @hmartenjoyer @CompleteSkeptic @justKDeng @mikeysee

看详情 →
sengpt@sengpt2026-09-17▲ 71.7k

arkadaşlar jev'e early access aldım ve test etmeye başladım ve gerçekten çok impressed olduğumu…

arkadaşlar jev'e early access aldım ve test etmeye başladım ve gerçekten çok impressed olduğumu söylemem lazım sadece şu kaydettiğim demoyu bi izleyin lütfen. kesinlikle hızlandırılmış bir video değil. 4 tane jev karşılıklı 101 oynuyor ve oyunun hızına bir bakın. beni takip edenler bilir. yıllardır her yeni model çıktığında bizim meşhur 101 okey oyununu oynatırım. benim kişisel benchmarkl

看详情 →
trou@_trou32026-09-17▲ 126.9k

For those who still didn't understand what Jev from can do, here is the example

For those who still didn't understand what Jev from @typesafeai can do, here is the example. Jev can read dozens of structured trading signals and then show intelligent decisions. It's time to make your system actually alive. What a time!

看详情 →
Sy@Sybuilds2026-09-17▲ 27.8k

Big AI news

Big AI news... Jev enables this microphone that mutes you when you talk BS @typesafeai

看详情 →
Pontus Karlsson@coolastekatten2026-09-17▲ 27.7k

Alright cooked Jev is truly the next era of building w LLMs Added this to Scappa tonight

Alright @typesafeai cooked Jev is truly the next era of building w LLMs Added this to Scappa tonight

看详情 →
Mormon Negro@mormonnegro2026-09-17▲ 25.8k

Acabo de terminar la implementación de + Chromium Headless para que mis agentes puedan navegar…

Acabo de terminar la implementación de @typesafeai + Chromium Headless para que mis agentes puedan navegar por internet a una buena velocidad! En este ejemplo le pido al agente que entre a la página del término "Café" en Wikipedia y navegue por los hipervínculos hasta terminar en la página "Inteligencia artificial" Logra completar la prueba en 20 segundos. Creo que pocos humanos podrían hacerlo

看详情 →
Isaac Tai 🧇@hi_im_isaac_2026-09-17▲ 665.7k

"jev can't generate text" But jev has a lot to say!

"jev can't generate text" But jev has a lot to say! I gave him a couple hundred common english words to choose from + punctuation

看详情 →
Neel Patel@Neel4902026-09-17▲ 336.7k

I think the implications of jev on self-driving could be huge

I think the implications of jev on self-driving could be huge

看详情 →
Shane@digitalshane_2026-09-17▲ 28.5k

I got access to Jev!

I got access to Jev! It can route website visitors to the pages they need based on what they need help with!

看详情 →
Lahfir@mdlahfir2026-09-17▲ 43.3k

Jev solved delegation

Jev solved delegation

看详情 →
Justin Schroeder@jpschroeder2026-09-16▲ 743.4k

I rebuilt Tesla Full Self Driving with Jev in less than an hour

I rebuilt Tesla Full Self Driving with Jev in less than an hour. This model is a total unlock.

看详情 →
Toran Bruce Richards@SigGravitas2026-09-16▲ 50.6k

JEV CAN DRIVE!

JEV CAN DRIVE! 🤯 I hooked up Jev to the raw controls of a driving simulator, this video shows him driving in realtime. Realtime is the kicker here. Asking a model what to do every 50ms then rendering that into a "realtime" video is one thing, but in this simulation I wanted to give Jev the challenge of actually controlling a moving vehicle in a simulation that will never pause while he thinks.

看详情 →
Raghav@RaghavPunnam2026-09-16▲ 65.8k

For those trying to understand the difference between Jev and LLMs + how to use it

For those trying to understand the difference between Jev and LLMs + how to use it

看详情 →
Lahfir@mdlahfir2026-09-16▲ 130.8k

Jev solved local harness/model routing I use a combination of Claude Code, Codex and Opencode…

Jev solved local harness/model routing I use a combination of Claude Code, Codex and Opencode as my local agentic stack and routing to other harnesses was always enforced in the system prompt/rules With a deterministic hook that Claude Code can decide before delegation, Jev helps to route to the right harness/model based on the task, and it's pretty accurate based on the intensity/intelligence o

看详情 →
it’s rivian@rrriviannn2026-09-16▲ 44.8k

I was wondering if Jev could generate images so here's what a sunflower looks like according to…

I was wondering if @typesafeai Jev could generate images so here's what a sunflower looks like according to Jev :

看详情 →
Anshu@anshuc2026-09-16▲ 401.3k

Jev can draw pictures by predicting every pixel in parallel

Jev can draw pictures by predicting every pixel in parallel

看详情 →
とさ@tosa_now2026-09-16▲ 41.8k

Jev ( ) 使って色々プロダクトで使えそうなの考えてみてる。

Jev (@typesafeai) 使って色々プロダクトで使えそうなの考えてみてる。 dndをしたら自動で仕分け先を教えてくれる体験 最初一回はちょっと遅いが、そのあとはサクサク

看详情 →
Justin Schroeder@jpschroeder2026-09-16▲ 329.8k

Silicon Valley predicted Jev a decade ago

Silicon Valley predicted Jev a decade ago.

看详情 →
cocktail peanut@cocktailpeanut2026-09-16▲ 167.8k

Silicon Valley predicted

Silicon Valley predicted @typesafeai

看详情 →
Faadil Shaik@faadilhshaik2026-09-16▲ 624.0k

got ’s new model Jev to play Super Mario Bros

got @typesafeai’s new model Jev to play Super Mario Bros. fast inference + structured outputs makes it surprisingly good for real time use cases. I'm excited to see what can be done with these new models!

看详情 →
Just_Lingonberry_352 🌼@JustLingonberry2026-09-16▲ 129.9k

playing minecraft, costing only 1 cent (150k tokens) for 2 minutes of play automatically…

@typesafeai playing minecraft, costing only 1 cent (150k tokens) for 2 minutes of play automatically knowing to dodge zombies at nightfall with no prompting

看详情 →
Yuntian Deng@yuntiandeng2026-09-16▲ 14.0k

I increasingly think future software will combine code for control flow with small neural…

I increasingly think future software will combine code for control flow with small neural programs for "fuzzy" judgments. That's what I've been exploring with ProgramAsWeights, which answers the question of where those neural programs come from: they are "compiled" from English descriptions. For example, I combined 30 neural programs with a decision tree to build a course website helper that loo

看详情 →
Teddo@Teddo_ICO2026-10-08▲ 104.6k

𝗕.𝗔𝗜 𝗕𝗿𝗶𝗻𝗴𝘀 𝗧𝘆𝗽𝗲𝗦𝗮𝗳𝗲 𝗝𝗲𝘃 𝗧𝗼 𝗔𝗜 𝗗𝗲𝗰𝗶𝘀𝗶𝗼𝗻 𝗠𝗮𝗸𝗶𝗻𝗴 What if an…

𝗕.𝗔𝗜 𝗕𝗿𝗶𝗻𝗴𝘀 𝗧𝘆𝗽𝗲𝗦𝗮𝗳𝗲 𝗝𝗲𝘃 𝗧𝗼 𝗔𝗜 𝗗𝗲𝗰𝗶𝘀𝗶𝗼𝗻 𝗠𝗮𝗸𝗶𝗻𝗴 What if an AI system could make structured decisions directly, instead of generating paragraphs that another system has to interpret? That is where Jev stands out. 𝗝𝗲𝘃 𝗕𝗿𝗶𝗻𝗴𝘀 https://t.c

看详情 →
Amar Harolikar@amarHarolikar2026-10-08▲ 96

The daily scan of the News headlines on the TREMOR app was a bit of a pain

The daily scan of the News headlines on the TREMOR app was a bit of a pain. It pulls about 800 headlines from 60+ sources. Now I hit Run Jev Rank, and in 2 seconds it scores and ranks the headlines based on my interest areas. The feeds come from news sites like Reuters, CNBC and

看详情 →
Grok@grok2026-10-08▲ 31

No. Laya uses RLCD (Reinforcement Learning for Calibrated Decisions) for training, the same…

@AtMemAi @atomic_chat_hq @OpenAI No. Laya uses RLCD (Reinforcement Learning for Calibrated Decisions) for training, the same method TypeSafe claims for Jev. Its architecture is a public bidirectional ModernBERT encoder plus option-marker decision head for single-pass non-autoregr

看详情 →
Grok@grok2026-10-08▲ 63

Jev came first

@AtMemAi @atomic_chat_hq @OpenAI Jev came first. TypeSafe AI released it on September 15, 2026. Laya followed three days later on September 18 as Convai’s open-weight response.

看详情 →
中山陽平 - 中小企業専門Webコンサル/650社/22年@b_gone2026-10-08▲ 186

「Web担当者向けSNS投稿リスクチェッカー」 Chrome ウェブストア Jevを使ってみたかったので、Chrome拡張機能リリースしました。選択して右クリックでポップアップします。

「Web担当者向けSNS投稿リスクチェッカー」 Chrome ウェブストアhttps://t.co/q0aUGs8E45 Jevを使ってみたかったので、Chrome拡張機能リリースしました。選択して右クリックでポップアップします。 ■概要 https://t.co/qcrOrh0zUs

看详情 →
Jatin Sood@Jatin_178242026-10-09▲ 43

The most talked-about AI model of 2026 can't write a single sentence

The most talked-about AI model of 2026 can't write a single sentence. Jev by TypeSafe AI doesn't chat. It decides, in under half a second. Will it replace LLMs? No. LLMs write. SLMs specialise. Jev decides. Full breakdown 👇 https://t.co/7YombJvuuu #JevAI #AI

看详情 →
Asaf Mazuz@MazuzAsaf2026-10-09▲ 169

Turn an open LLM into Jev-shaped decisions

Turn an open LLM into Jev-shaped decisions. Nokia research. No TypeSafe key on the local path. ANYJEV Choice, Noul, and Score with probabilities, from your own model. Training-free first, then optional calibration. (~1,030★) When to use: typed decisions on open weights you

看详情 →
HooHa@byubyv2026-10-10▲ 4

一个不生成文字的 AI,开发商估值已到 75 亿美元。据 TechCrunch,TypeSafe AI 融资 8.7 亿美元,a16z 领投。它的 Jev 输出概率,面向自动化决策。AI…

一个不生成文字的 AI,开发商估值已到 75 亿美元。据 TechCrunch,TypeSafe AI 融资 8.7 亿美元,a16z 领投。它的 Jev 输出概率,面向自动化决策。AI 产品的竞争正在扩展到一个新问题:怎样给机器一个可用的判断。 https://t.co/nKs40BOh3O

看详情 →
Naveen@NaveenKumar__T2026-10-10▲ 41

TypeSafe AI raised $870M at a $7.5B valuation (a16z lead) for Jev, a transformer that outputs…

TypeSafe AI raised $870M at a $7.5B valuation (a16z lead) for Jev, a transformer that outputs calibrated decisions, not text. Launched Sept 15; claims 1/3 of Fortune 500 use it (TechCrunch). Not every AI needs to talk. Some just need to decide. #AI #Startups #Automation

看详情 →
AwesomeAI@Awesome_AI_News2026-10-10▲ 25

TypeSafe AI launched Jev, a new AI decision model, creating a stir within weeks

TypeSafe AI launched Jev, a new AI decision model, creating a stir within weeks. It raised $870M led by Andreessen Horowitz with Sequoia Capital and angels, reaching a $7.5B valuation. Jev gained 1M users in days and is used by about one-third of Fortune 500 companies.....

看详情 →
Inside Frontier@insidefrontierV2026-10-10▲ 22

Jev 如何让 AI 的回答变得可检查?从有限选项、并行评估到显式概率,这个中文视频讲清 Choice、Score 和 Noul 的工作方式,也解释为什么类型安全仍然需要语义验证。

Jev 如何让 AI 的回答变得可检查?从有限选项、并行评估到显式概率,这个中文视频讲清 Choice、Score 和 Noul 的工作方式,也解释为什么类型安全仍然需要语义验证。 #Jev #AIAgents https://t.co/dDMnQ7ocHr https://t.co/G1sXbJd3hC

看详情 →
Bevan Ho@ho_bevan2026-10-10▲ 39

前 OpenAI 研究員推 Jev,專攻結構化回應,讓 AI 更穩定可預期。對依賴輸出格式的應用很有幫助。

前 OpenAI 研究員推 Jev,專攻結構化回應,讓 AI 更穩定可預期。對依賴輸出格式的應用很有幫助。 #AI #TypeSafe

看详情 →
cloud99 🚀 AI&Electronic@99ninecloud2026-10-10▲ 24

非文本AI模型Jev开发商TypeSafe AI估值75亿美元 TypeSafe AI完成8.7亿美元融资,估值75亿美元,由a16z领投,红杉及老股东DCVC参投。Jev于9月15日发布,不…

非文本AI模型Jev开发商TypeSafe AI估值75亿美元 TypeSafe AI完成8.7亿美元融资,估值75亿美元,由a16z领投,红杉及老股东DCVC参投。Jev于9月15日发布,不输出文字,而是输出概率形式的“校准后的决策”;公司称已有三分之一的财富500强企业在使用。以上仅为事实整理,不构成投资建议。

看详情 →
缪斯观察@martin520_xyz2026-10-10▲ 22

TypeSafe AI 的 Jev 发布仅三周多,就以 75 亿美元估值融资 8.7 亿美元,a16z 领投,红杉跟投。它不输出文字,而是输出校准后的概率判断,号称更快、更省…

TypeSafe AI 的 Jev 发布仅三周多,就以 75 亿美元估值融资 8.7 亿美元,a16z 领投,红杉跟投。它不输出文字,而是输出校准后的概率判断,号称更快、更省 token,已被三分之一的财富 500 强使用。LLM 之外,"决策模型"成了新赛道。

看详情 →
Sina 🗝️⚡ BI Report@Snz_BTC2026-10-10▲ 3.8k

Days after the new class of LLMs, JEV, came along, openAI quickly introduced their own decision…

Days after the new class of LLMs, JEV, came along, openAI quickly introduced their own decision model. A couple days later, Microsoft. Typesafe called JEV a System 1 model. Open AI pretended they are not copying them and went with completely different name: decision model. https:

看详情 →
Eric Pereira@iamericpereira2026-10-11▲ 205

TypeSafe AI raised $870M at a $7.5B valuation this week

TypeSafe AI raised $870M at a $7.5B valuation this week. 24 days earlier they were a $40M seed nobody had heard of. The company is real and the numbers are real (TechCrunch, Bloomberg, Crunchbase all confirm). But the money did not arrive because of the model. It arrived because

看详情 →
Kuman@KUMAN_R2026-10-10▲ 25.1k

シリーズ A でいきなりユニコーンか.....🦄。Jev を出した TypeSafe が、公開から3週間ほどで a16z リードのシリーズ A を発表、、 ・調達額 —…

シリーズ A でいきなりユニコーンか.....🦄。Jev を出した TypeSafe が、公開から3週間ほどで a16z リードのシリーズ A を発表、、 ・調達額 — 8億7,000万ドル(約1,380億円) ・評価額 — 75億ドル(約1.2兆円) ・顧客 — Fortune 500 の3分の1がすでに利用 https://t.co/LRtNkbqfas

看详情 →
Open Loop News | AI News@OpenLoopNews2026-10-10▲ 4

TypeSafe AI raised $870M at a $7.5B valuation for Jev, an AI model that gives answers but…

TypeSafe AI raised $870M at a $7.5B valuation for Jev, an AI model that gives answers but writes no text. https://t.co/NPVWg8q3LH

看详情 →
The Hallway Track@TheHallwayTrk2026-10-10▲ 9

Decision models just became a product category

Decision models just became a product category. 🔥 OpenAI, Microsoft, Perplexity, Cloudflare, Liquid all shipped on the same day. TypeSafe/Jev hit $100M ARR in a week. Typed, single-pass outputs are the new AI primitive. 📌 #LLMs #TheHallwayTrack

看详情 →
Ali Nasser@_alithegreat_2026-10-11▲ 40

TypeSafe AI reportedly raised $870M in a Series A at a $7.5B valuation

TypeSafe AI reportedly raised $870M in a Series A at a $7.5B valuation. Its Jev model focuses on calibrated decisions, not conventional text generation. Could specialized #AI become essential software infrastructure—or will the valuation be hard to justify?

看详情 →
MarsandMoon@MarsandMoon20262026-10-11▲ 26

AI路线战争开打:不比谁会聊天,比谁敢下判断。

AI路线战争开打:不比谁会聊天,比谁敢下判断。 TypeSafe以75亿美元估值融资8.7亿,a16z领投——模型Jev不写作文,只输出带置信度的决策概率。 第一性原理:商业的稀缺品从来不是观点,是校准过的判断。 流量属于聊天框,利润属于敢为概率负责的人。

看详情 →
Mehdi Allahyari@MehdiAllahyari2026-10-11▲ 4.2k

استارتاپ TypeSafe AI که پشت مدل Jev قرار دارد، خبر جذب سرمایه جدیدش را به سبک متفاوتی با نام…

استارتاپ TypeSafe AI که پشت مدل Jev قرار دارد، خبر جذب سرمایه جدیدش را به سبک متفاوتی با نام Series AI منتشر کرده است. به جای تب و تاب معمول چت‌بات‌ها، تمرکز آنها روی مدل‌های به اصطلاح ماشین‌محور (machine-native) یا مدل‌های سیستم یک (System One) است؛ مدل‌هایی که هدفشان تولید متن

看详情 →
Yeet Capital@yeeetCapital2026-10-12▲ 61

JUST IN: TypeSafe AI raises an $870M Series A at a $7.5B valuation, led by Andreessen Horowitz…

JUST IN: TypeSafe AI raises an $870M Series A at a $7.5B valuation, led by Andreessen Horowitz, three weeks after launching its Jev decision model.

看详情 →
STARTUP LOG|スタートアップの挑戦を、もっと身近に| News Media|スタログ@STARTUPLOG_COM2026-10-13▲ 195

ソフトウェアの意思決定を支援するAIモデル「Jev」を開発するTypeSafe AIが8億7,000万ドルの資金調達を実施

ソフトウェアの意思決定を支援するAIモデル「Jev」を開発するTypeSafe AIが8億7,000万ドルの資金調達を実施 https://t.co/xXjBH5d4M9

看详情 →
致龙@wanggeng6662026-10-13▲ 16

上周,以太坊研究界的一位知名人士发表了一份令人震惊的声明,暗示人工智能可能在几个月内破解 ECDSA,而 ECDSA 是整个行业所依赖的公钥加密算法。

上周,以太坊研究界的一位知名人士发表了一份令人震惊的声明,暗示人工智能可能在几个月内破解 ECDSA,而 ECDSA 是整个行业所依赖的公钥加密算法。 据说只订着50个以上比特币的钱包 一场新的人工智能竞赛正在迅速兴起,微软(MSFT)也以惊人的姿态加入了这场竞赛。 在 TypeSafe 发布 Jev

看详情 →
Matt Green@IAmMattGreen2026-10-13▲ 282

typesafe is hiring a "safety jevroni" to help align Jev

typesafe is hiring a "safety jevroni" to help align Jev https://t.co/kJzL0bOUMO

看详情 →
後藤康成 | AIとラグビーニュースのアンパサンド CEO@got2026-10-13▲ 103

Microsoftは文章生成ではなく、提示した選択肢ごとの確率を返す判断特化型モデル「Microsoft-Decision-1」を公開した。AlibabaのオープンウェイトモデルQwen3.5…

Microsoftは文章生成ではなく、提示した選択肢ごとの確率を返す判断特化型モデル「Microsoft-Decision-1」を公開した。AlibabaのオープンウェイトモデルQwen3.5-9Bを追加学習し、状況、質問、固定回答候補を受けてJSON形式で確率を出力する。料金は入力100万トークン0.042ドル、出力無料で、TypeSafe

看详情 →
とむ|SaaS事業企画|AI×GTM@tomaigtm2026-10-13▲ 99

会話をしないAIが、3週間で評価額約75億ドル(約1.1兆円)になった。

会話をしないAIが、3週間で評価額約75億ドル(約1.1兆円)になった。 TypeSafe AIのJevというモデルは、チャットのように文章を作るのではなく、「はい/いいえ」や「この中から選ぶ」といった決定を瞬時に出す専用モデル。 https://t.co/eIUp6871Pd

看详情 →
你是来拉屎的吧?@lucky94w2026-10-13▲ 86

OpenAI在自研绞杀对手这件事上下手真的快。前OpenAI核心员工Diogo Almeida离职创办TypeSafe并推出Jev,上线才15天就传出按超100亿美元估值融资,主打的就是不做长…

OpenAI在自研绞杀对手这件事上下手真的快。前OpenAI核心员工Diogo Almeida离职创办TypeSafe并推出Jev,上线才15天就传出按超100亿美元估值融资,主打的就是不做长文本聊天、专门针对选项做超低延迟结构化选择。结果山姆奥特曼在今天的DevDay上直接掏出Decisions

看详情 →
Muse Talks@MuseTalks2026-10-14▲ 7

typesafe raised $870m for jev, a model that never talks to you

typesafe raised $870m for jev, a model that never talks to you. no sentences, no chat. it outputs calibrated decisions straight into the software. a third of the fortune 500 reportedly use it already. two years grading models on words. this one gets graded on its error bars. http

看详情 →
갑오징어@koreabuldak2026-10-14▲ 478

고르고 점수 매기는 일만 하는 AI '의사결정 모델' JEV-35B가 허깅페이스 트렌딩에 올라왔길래 봤는데, 내가 맥미니 로컬 모델로 하는 기사 점수 매기기와 같은 종류의…

고르고 점수 매기는 일만 하는 AI '의사결정 모델' JEV-35B가 허깅페이스 트렌딩에 올라왔길래 봤는데, 내가 맥미니 로컬 모델로 하는 기사 점수 매기기와 같은 종류의 일을 하는 모델이었음. 모델 카드로는 AutoTrust AI가 Qwen3.5-35B-A3B를 바탕으로 만든 오픈 모델(아파치 2.0)임. 글은 안 쓰고 https://t.co/amdUiyaOnq

看详情 →
新井 元気@ITエンジニア@_moto___2026-10-13▲ 802

「判定専用AI『Jev』をエージェント基盤に組み込んでみた」の記事をまとめてみた。前提知識も説明してあるので読んでみてほしい。

「判定専用AI『Jev』をエージェント基盤に組み込んでみた」の記事をまとめてみた。前提知識も説明してあるので読んでみてほしい。 --- ■先に押さえる3つのポイント 1. 正解率は同じで、判定コストは約150分の1になったこと。 2. https://t.co/fMgAQSXyUd https://t.co/pQBGfkVK8Z

看详情 →
Everymans.ai@EverymansAI2026-10-13▲ 140

🔸S1 Watch —DeepEval introduces JevEval uses ➞ TypeSafe AI: Decision-scoring models just became…

🔸S1 Watch #3—DeepEval introduces JevEval uses ➞ TypeSafe AI: Decision-scoring models just became judges. Judges don't release answers.💥 🧐 S1 Watch Series: #1 https://t.co/ETLMZnKmjD™ + Jev integration post: https://t.co/r2WSMsrZYV #2 @Microsoft & @OpenAI competing with htt

看详情 →
はち@CurveWeb2026-10-13▲ 11.4k

Jevライクなモデル(Decision Model)の強化学習を試しました。

Jevライクなモデル(Decision Model)の強化学習を試しました。 TypeSafe AI はJevを強化学習したと述べていますが、具体的なやり方は公開されておらず、コミュニティでも今の所これという方法は定まっていません。 1つの例として見ていただければと思います。 https://t.co/x1LHvrDo81

看详情 →
Tatsuya@tatsuyahobby2026-10-15▲ 1

TypeSafe AI のJevの話題がTLに流れてこなくなって(ブームが去った?)から、招待メール…コレどうなんでしょう?遅えよ。出遅れた感半端ねえ。他の人がやりつくした感ある後に来るとかモ…

TypeSafe AI のJevの話題がTLに流れてこなくなって(ブームが去った?)から、招待メール…コレどうなんでしょう?遅えよ。出遅れた感半端ねえ。他の人がやりつくした感ある後に来るとかモチベーションがダダ下がり。

看详情 →
StudioYebisu@studio_yebisu2026-10-15▲ 5.9k

【悲報】Jev本家こと、TypeSafeさん。

【悲報】Jev本家こと、TypeSafeさん。 たった今、画像に対応するもすでにありとあらゆるJevクローンがローンチされているので。 あまり話題にならない。でも、わしは本家の世界知識の多さを評価してるで。 https://t.co/tv6k0fxUXD

看详情 →
Dot@Dot202609302026-10-15▲ 11

🖼️ 框得出来,也可能根本没目标。

🖼️ 框得出来,也可能根本没目标。 TypeSafe 10月14日发布 Jev 1.14,新增单个边界框输出。官方文档提醒:目标缺席仍会返回框,需搭配 Noul 判断是否存在。 ⚠️ 图像功能仍属实验。自动裁图前,别忘了测“空目标”的误裁率。 #AI #多模态 https://t.co/seXtJ5QXuA

看详情 →
Jay Zhou@jayzhoupro2026-10-15▲ 83

Holy crap! 👀 Jev can take images inside the same structured request you already use for…

Holy crap! 👀 Jev can take images inside the same structured request you already use for decisions. @CompleteSkeptic's release thread opens up experimental image support. TypeSafe's blog says images can sit in state, instructions or criteria. The new box type returns a single htt

看详情 →
スナガク | Codexではじめるエージェンティックコーディング@suna_gaku2026-10-15▲ 270

① Jev 1.14.0 では多言語対応を中心に改善 大きな新機能よりも、モデル全体の細かな品質向上を重視。特に多言語対応での改善が期待されている。まず jev-preview…

① Jev 1.14.0 では多言語対応を中心に改善 大きな新機能よりも、モデル全体の細かな品質向上を重視。特に多言語対応での改善が期待されている。まず jev-preview で公開され、約 1 週間後に jev-latest へ反映する予定。 ② 画像を使った判断が可能になった Jev

看详情 →
WquGuru@wquguru2026-10-16▲ 9.6k

🎉 Jev 1.14 终于能看图了!!!

🎉 Jev 1.14 终于能看图了!!! 我让它在一幅包含上百个小人的画里找一个人,看看效果如何。 尼德兰箴言里藏了上百人物和动物,这里把找人拆成三个步骤: 1. 选块:把画切成 16 块,出一道 choice,16 个选项本身就是 16 张图。它选第 12 块,概率为 0.95 2. 框:在这一块上问一道 https://t.co/5o4B0ptUp4 https://t.co/H3vVMZ5x18

看详情 →
˗ˋˏ コグレマサト|ネタフル ˎˊ˗@kogure2026-10-16▲ 392

“1つ1つの記事タイトルと公開日付から「今すぐ読むべきか(Noul)」「今日中に読む価値がどれくらいあるか(Score)」「記事ジャンルは何か(Choice)」を判断してもらう、といった使い方…

“1つ1つの記事タイトルと公開日付から「今すぐ読むべきか(Noul)」「今日中に読む価値がどれくらいあるか(Score)」「記事ジャンルは何か(Choice)」を判断してもらう、といった使い方が考えられます。” / “日沼諭史の「AI活用入門塾」 判断だけができるAI「Jev」、使い…” https://t.co/iuTxRB4gzd

看详情 →
Nathan Khoo@natebakescakes2026-10-16▲ 112

Jev can see now

Jev can see now. According to TypeSafe's release notes, the last version could see too and just never mentioned it. Jev 1.14 adds three things to play with. You can put an image anywhere you'd normally put text. There's a new question type called Box: ask where something is in a

看详情 →
いにしえ@高信頼AIニュース"NeuralWire.org"運営|Will Oldgram@old_pgmrs_will2026-10-16▲ 609

Jevが画像にも対応 そういえば TypeSafe AI は直近で大規模な資金調達にも成功してた 類似モデルが続出してる状況で今後の伸びに期待できるかな?

Jevが画像にも対応 そういえば TypeSafe AI は直近で大規模な資金調達にも成功してた 類似モデルが続出してる状況で今後の伸びに期待できるかな? https://t.co/rRpZcNnB1S https://t.co/CO0wjYGu3V

看详情 →
株式会社FLINTERS💻@flinters_inc2026-10-15▲ 216

📣お役立ち記事公開📖 FLINTERS LABで、AIモデル「Jev」の解説記事を公開しました。

📣お役立ち記事公開📖 FLINTERS LABで、AIモデル「Jev」の解説記事を公開しました。 「判断が速い」と話題のJevの正体とLLMとの違い、試し方を、FLINTERSでの検証事例とあわせてまとめています。 業務効率化のヒントとして、ぜひご覧ください👇 https://t.co/5lZUmYxyrB #Jev #TypeSafe #生成AI

看详情 →
ねじまきちゃん@nejimaki_ai2026-10-15▲ 129

文章を書かず、「AかBか」の判断だけを確率つきで返すAI「Jev」(TypeSafe AI)は、9月の公開時は文字しか受け取れず、写真を判断に使うには先に文章へ書き起こす必要がありました。10…

文章を書かず、「AかBか」の判断だけを確率つきで返すAI「Jev」(TypeSafe AI)は、9月の公開時は文字しか受け取れず、写真を判断に使うには先に文章へ書き起こす必要がありました。10月14日のバージョン1.14から、画像をそのまま渡せるようになりました。 【画像対応でできるようになったこと】

看详情 →
ねじまきちゃん@nejimaki_ai2026-10-15▲ 10

TypeSafe AI の Jev(文章を書かず、選択肢ごとの確率だけを返す判断専用モデル)が 1.14 で画像を受け取れるようになりました。公式ドキュメントを読むと、位置を返す新しい質問の型…

TypeSafe AI の Jev(文章を書かず、選択肢ごとの確率だけを返す判断専用モデル)が 1.14 で画像を受け取れるようになりました。公式ドキュメントを読むと、位置を返す新しい質問の型 Box には「見つからない」という答えがありません。写っていない物を聞いても、枠が返ってきます。

看详情 →
梦莱加德@EuneJjj738572026-10-17▲ 111

$6,240 本金入场,BTC 结算拉升至 $13,180。

$6,240 本金入场,BTC 结算拉升至 $13,180。 本场最大单笔盈利的建仓,仅耗时 184 毫秒。 ​这段实盘录屏来自 grokbot + jev 组合。 六个 grok 节点 一只粉色蜘蛛爬虫 jev 盯盘 5 分钟级别 实际执行在 1 分钟线。 ​蜘蛛抓数据,jev 下判断,grok 节点直接开火。 ​"breakout confirmed" - 捕捉突破

看详情 →
Tesla_Optimus@Tesla_Optimus_K2026-10-17▲ 65

Jev just became Vercel's fastest-adopted model ever Two weeks after a limited launch on…

Jev just became Vercel's fastest-adopted model ever Two weeks after a limited launch on September 15, TypeSafe AI's Jev is on 35% of Vercel's paid teams — nearly double Sonnet 5.5, the next most-used model in that window. Jev is a small decision model: pick among a few preset htt

看详情 →
後藤康成 | AIとラグビーニュースのアンパサンド CEO@got2026-10-17▲ 80

TypeSafe AIの意思決定モデルJevが、限定公開から2週間でVercelの有料チームの35%に使われ、同社AI…

TypeSafe AIの意思決定モデルJevが、限定公開から2週間でVercelの有料チームの35%に使われ、同社AI Gatewayで過去最速の初期採用を記録した。利用範囲は2位のSonnet

看详情 →
CodingAfterDark@CodingAfterDrk2026-10-17▲ 20

Jev, the AI that picks instead of writes: what it’s good for, and the catch Most AI writes

Jev, the AI that picks instead of writes: what it’s good for, and the catch Most AI writes. Jev, a new model from TypeSafe AI, only decides, and it does it in about a tenth of a second for a fraction of a cent. https://t.co/NZYmrFjneQ #codingafterdark #AI

看详情 →
akuwano@kuwa_tw2026-10-17▲ 76

いえ、恐らくUnity Gatewayを通して統一したガバナンスでTypeSafeさん側のJevを使えるよって話だと思いますねー!

@eiraces いえ、恐らくUnity Gatewayを通して統一したガバナンスでTypeSafeさん側のJevを使えるよって話だと思いますねー!

看详情 →
Zeka Hattı@ZekaHatti2026-10-17▲ 7

Vercel'e göre TypeSafe AI'ın hafif "karar modeli" Jev, platformun en hızlı benimsenen YZ modeli…

Vercel'e göre TypeSafe AI'ın hafif "karar modeli" Jev, platformun en hızlı benimsenen YZ modeli oldu: Çıkıştan iki hafta sonra ücretli ekiplerin %35'i kullanıyordu; bu, ikinci sıradaki Sonnet 5.5'in neredeyse iki katı. Kaynak: TechCrunch

看详情 →
akuwano@kuwa_tw2026-10-17▲ 1.9k

DatabricksでTypeSafeさんの意思決定モデルである「Jev-1.14」が利用可能になりました!

DatabricksでTypeSafeさんの意思決定モデルである「Jev-1.14」が利用可能になりました! 分類やスコアリングに特化、標準的LLMより最大200倍高速かつ400倍低コストで、大規模データをリアルタイムに処理可能! ネイティブ機能であるai_decideもありますが、本家のJevも使えるようになりました! https://t.co/HC0igjiM2v

看详情 →
VerySmallWoods@verysmallwoods2026-10-17▲ 388

的 Jev 更新到了 1.14。

@typesafeai 的 Jev 更新到了 1.14。 官方列出的变化包括: - 多项小幅改善和多语言能力提升 - 新增了 Box 回答类型 Box 类型可以用上、下、左、右四条边表示目标在图片中的位置。 我通过 TypeSafe 官方 https://t.co/IBCXyQuwj8

看详情 →
しろちゃん@AI情報@shirochange_ai2026-10-19▲ 48

【海外ニュース】TypeSafe AIの軽量モデル「Jev」が、Vercelで公開2週間で有料チームの35%に採用され同社史上最速の普及に。決まった選択肢から「判断だけ」を行う設計で、大量デー…

【海外ニュース】TypeSafe AIの軽量モデル「Jev」が、Vercelで公開2週間で有料チームの35%に採用され同社史上最速の普及に。決まった選択肢から「判断だけ」を行う設計で、大量データの分類を高速・低コストに。(TechCrunch) #Jev #生成AI https://t.co/gtDpF5I6ue

看详情 →
Neovek@neovek_ai2026-10-19▲ 25

Le modèle de décision Jev rend des choix typés avec probabilités, sans texte

Le modèle de décision Jev rend des choix typés avec probabilités, sans texte. Un LLM à sortie structurée donne déjà ce format. Ce qui change : le prix, la latence et la probabilité. Encore faut-il qu'elle soit calibrée, ce que TypeSafe n'a pas montré. https://t.co/aq2iBnwbWx

看详情 →
生成AIでIPチャレンジ@GenAI_Creative2026-10-19▲ 133

約1375億円調達、新興TypeSafe AIの予測AI「Jev」は導入企業に利益を生むか | Forbes JAPAN 公式サイト(フォーブス ジャパン)

約1375億円調達、新興TypeSafe AIの予測AI「Jev」は導入企業に利益を生むか | Forbes JAPAN 公式サイト(フォーブス ジャパン) https://t.co/G1hsa88j11

看详情 →
율무커피 YulmuCoffee@yulmu_coffee2026-10-25▲ 3.8k

TypeSafe의 Jev

TypeSafe의 Jev.... Noul 출력물(예/아니오 확률 반환)에 대해 대규모 유출이 있었다고 합니다. https://t.co/WkNM8XX1qD

看详情 →
MLS X 🗽@MLS_HQ2026-10-24▲ 74

SITUACIÓN: TypeSafe AI, creadora del modelo Jev, admite que se filtraron todas las salidas de…

SITUACIÓN: TypeSafe AI, creadora del modelo Jev, admite que se filtraron todas las salidas de sus clientes; el CEO Diogo Almeida lo reveló en X y publicó el enlace a los datos. Sin causa oficial confirmada. Fuente: Startup Fortune https://t.co/eNgN2khWAx

看详情 →
Diego Turati@diego_turati2026-10-24▲ 58

Meet the TypeSafe AI connector in Integration Service

Meet the TypeSafe AI connector in Integration Service. 🔗 Bring Jev models into your workflows to make fast, confident decisions at lower cost & latency: https://t.co/SpNk7GAK6n https://t.co/dSeBF8G3lv

看详情 →
Arian Ramirez@ari4n122026-10-25▲ 11

El contexto: • Jev se lanzó a mediados de septiembre • TypeSafe había levantado $40M (The…

El contexto: • Jev se lanzó a mediados de septiembre • TypeSafe había levantado $40M (The Register) • Promete respuestas en 70–500 ms, mucho más rápido y barato que un LLM Velocidad y precio vendieron. La seguridad no escaló igual de rápido.

看详情 →
Arian Ramirez@ari4n122026-10-25▲ 13

🚨 Una startup de IA con apenas un mes en el mercado acaba de admitir que se filtraron TODAS…

🚨 Una startup de IA con apenas un mes en el mercado acaba de admitir que se filtraron TODAS las salidas que su modelo generó para sus clientes. Se llama TypeSafe AI. Su modelo: Jev. La lección para founders es enorme 🧵👇 https://t.co/CuEtynaimL

看详情 →
Mission Ready@MissionReadyQA2026-10-25▲ 3

Meet the TypeSafe AI connector in Integration Service

Meet the TypeSafe AI connector in Integration Service. 🔗 Bring Jev models into your workflows to make fast, confident decisions at lower cost & latency: https://t.co/GV7yBJ8qh4 https://t.co/jgEjXGvTs8

看详情 →
agentictesting@agentictesting2026-10-25▲ 35

Meet the TypeSafe AI connector in Integration Service

Meet the TypeSafe AI connector in Integration Service. 🔗 Bring Jev models into your workflows to make fast, confident decisions at lower cost & latency: https://t.co/wsCmmNHIgc https://t.co/mYmCoP72JG

看详情 →
mike buttrum@mikebuttrum2026-10-25▲ 44

Meet the TypeSafe AI connector in Integration Service

Meet the TypeSafe AI connector in Integration Service. 🔗 Bring Jev models into your workflows to make fast, confident decisions at lower cost & latency: https://t.co/Y1L7q4VZAN https://t.co/jdV0tttS4x

看详情 →
robots4testing@robots4testing2026-10-25▲ 35

Meet the TypeSafe AI connector in Integration Service

Meet the TypeSafe AI connector in Integration Service. 🔗 Bring Jev models into your workflows to make fast, confident decisions at lower cost & latency: https://t.co/TYMNMvJnI8 https://t.co/Sf4w7gv9rr

看详情 →
Ben@brochchain2026-10-25▲ 89

Jev 1.14 from TypeSafe AI can now take pictures, point to where things are in them, and read…

Jev 1.14 from TypeSafe AI can now take pictures, point to where things are in them, and read about 400 pages in one question. https://t.co/Y1xbkcw1Qa

看详情 →
reijjb@reijjb2026-10-26▲ 102

TypeSafe suspended new registrations on Jev on September 22, 2026, because after the full…

TypeSafe suspended new registrations on Jev on September 22, 2026, because after the full opening of access on the 20th, such a wave of demand swept in that the service simply did not drag on. Officially: "immense swell of demand" — it was necessary to protect the quality for htt

看详情 →
Webbovich@webbovich772026-10-26▲ 67

Similar to Doomers, which Typesafe (Jev), Higgsfield and other use

@elshayib_ Similar to Doomers, which Typesafe (Jev), Higgsfield and other use https://t.co/No8CQFY0zS

看详情 →
Archimedes@ArchimedesLore2026-10-28▲ 1.7k

WHOEVER LEAKED THIS JEV CASCADE SHOULD BE REPLACING FINE-TUNING I thought teaching a model…

WHOEVER LEAKED THIS JEV CASCADE SHOULD BE REPLACING FINE-TUNING I thought teaching a model which bucket to pick meant 10,000 labeled examples and a two-day training run on H100s. Apparently not. This cascade's setup is much more interesting: → CHOICE accepts a criteria dict https

看详情 →
Soups Ranjan@soupsranjan2026-10-28▲ 1.3k

Next week we’re co-hosting an event with and called ‘Jev and the Future of Fraud Fighting.’…

Next week we’re co-hosting an event with @a16z and @typesafeai called ‘Jev and the Future of Fraud Fighting.’ When Jev came out a month ago, I tested it on an actual fraud incident at a Sardine customer. What I found shocked me: - Jev outperformed LLMs in fraud applications and h

看详情 →
NOOR TECH@NoorTech02026-10-28▲ 30

TypeSafe makes Jev learning delightful

@typesafeai @allietheicon TypeSafe makes Jev learning delightful

看详情 →
Rodrigo Uroz@rodrigouroz2026-10-28▲ 36

Then Jev from TypeSafe decides if it matters: family, school, payments, bills

Then Jev from TypeSafe decides if it matters: family, school, payments, bills. DMs always wake Eco. Groups only if the score is high. If Jev fails, Eco wakes up anyway. I'd rather get a false alarm than miss something.

看详情 →
𝓑𝓲𝓬_𝓡𝓮𝓿𝓮𝓵𝓪𝓽𝓲𝓸𝓷@Bic_Revelation2026-10-31▲ 472

There's a smarter way to cut AI costs than switching to a cheaper LLM

There's a smarter way to cut AI costs than switching to a cheaper LLM. Stop sending every decision to an LLM in the first place. I'd build a three-layer intelligence system using: Jev + DuckDB + Claude + jev-calibrate. The idea is simple. Let software handle facts, Jev https://t.

看详情 →
中村 亮一 @投資会社経営 | ベンチャー投資 | 上場株 | 収益不動産 | 事業再生@nack8232026-10-31▲ 158

【 #気になった案件 】 ↓この企業というかビジネスが、めっちゃ気になる TypeSafe AI 企業概要:

【 #気になった案件 】 ↓この企業というかビジネスが、めっちゃ気になる TypeSafe AI 企業概要:

看详情 →
奶茶不加冰@wangxiaoxin882026-10-31▲ 673

Google 搜不到自家站点,这事儿我太懂了——上线几个月连个影儿都没有,先别掏钱请人诊断,也别抱着官方文档硬啃、不知道从哪下刀。

Google 搜不到自家站点,这事儿我太懂了——上线几个月连个影儿都没有,先别掏钱请人诊断,也别抱着官方文档硬啃、不知道从哪下刀。 后来我盯上了 jev-seo:给个首页网址,大约一分钟就能把整站爬完,再按 Google 官方搜索文档那 52 https://t.co/IDzmyhZ8JR

看详情 →
PYXL@pyxl_dog2026-10-30▲ 47

Mod Claude Code : avant chaque message, Jev (TypeSafe, via OpenRouter) choisit l’effort de…

Mod Claude Code : avant chaque message, Jev (TypeSafe, via OpenRouter) choisit l’effort de réflexion : low, medium ou high. Le modèle ne change pas, le cache reste intact. Moins d’1 s par message. https://t.co/Cr4FM3SHKQ

看详情 →
kekeyon3@ワサラー団 ワサラー団十六夜 ❄❄❄❄❄❄@kekeyon32026-10-30▲ 101

Jevとは何か。TypeSafeが出している「System One」系の意思決定専用モデル。

Jevとは何か。TypeSafeが出している「System One」系の意思決定専用モデル。 普通のLLMのように文章を生成しない。 状態(目的・エラー・試行回数など)を渡す 選択肢の中から確率付きで答える 入力 $0.042 / 100万トークン、出力無料 数百ミリ秒程度で返ってくる

看详情 →
Milton Smith@javamuffinztx2026-11-01▲ 76

Added TypeSafe AI's Jev system one model to my app for Ai model routing and a few other…

Added TypeSafe AI's Jev system one model to my app for Ai model routing and a few other features. I don't have a big log of activity, as you can see lol, but it's a 38x increase in performance over thinking models (my fall back). I'm thinking it would probably make a good tool ht

看详情 →
一田和樹@K_Ichida2026-11-01▲ 294

「判断モデル」Jevの誇大宣伝と短期間での失速を技術者が分析。

「判断モデル」Jevの誇大宣伝と短期間での失速を技術者が分析。 TypeSafe AIが「System One

看详情 →
Toshifumi Kuga@ToshifumiKuga2026-11-01▲ 136

On October 9, 2026, Microsoft released a new decision AI model

On October 9, 2026, Microsoft released a new decision AI model. It is named "https://t.co/D1gIfzMrao this article, I will test its performance by comparing it with several other decision AI models, including its predecessor, Jev Hope you enjoy this. https://t.co/DK79umoO2H

看详情 →
Artur Nowocin@arturnow2026-10-31▲ 217

Every model talks

Every model talks. This one doesn't — it just decides. 70 milliseconds, honest probabilities, 400x cheaper. The most important AI of the next few years might be the one you never talk to. https://t.co/asajEB8mUb

看详情 →
Shingo|AIスタートアップ→FDE@japan198408242026-11-02▲ 1.5k

文章を書かず判定だけ返すAI「Jev」で、国会の質問2,424件を仕分けたら約45円でした。規制緩和を事業者が早く知る方法を探し、先読みは外れ、見送りの判定は正解率が約5割。国の法律と福井県の…

文章を書かず判定だけ返すAI「Jev」で、国会の質問2,424件を仕分けたら約45円でした。規制緩和を事業者が早く知る方法を探し、先読みは外れ、見送りの判定は正解率が約5割。国の法律と福井県の予算を業種で探せるサイトになるまでの約40日の記録です。 #生成AI #規制緩和 https://t.co/j9HIpQIKrU

看详情 →
つれづれ備忘録@tsurezure_memo2026-11-01▲ 0

TypeSafe AIのJevは文章を作らず、あらかじめ決めた選択肢や点数の形で判定を返します。生成と識別の軸で見ると、生成よりも識別に近い種類です。ただ、AIの分け方そのものに、決まった1つ…

TypeSafe AIのJevは文章を作らず、あらかじめ決めた選択肢や点数の形で判定を返します。生成と識別の軸で見ると、生成よりも識別に近い種類です。ただ、AIの分け方そのものに、決まった1つの答えはありません。 https://t.co/sdUfG5Jtzr

看详情 →
まにらン新聞@ManiRanShinbun2026-11-01▲ 60

フィリピンで育ったDiogo Almeida氏が共同創業した米サンフランシスコの人工知能(AI)新興企業TypeSafe…

フィリピンで育ったDiogo Almeida氏が共同創業した米サンフランシスコの人工知能(AI)新興企業TypeSafe AIが、AIモデル「Jev」で有力投資家の資金と国際的な注目を集めている。同氏はフィリピン人とポルトガル人の血を引き、同社の最高経営責任者(CEO)を務める。

看详情 →
Grok@grok2026-11-01▲ 32

The core pattern holds up

@AndresUribeRen1 @xsmryze The core pattern holds up. Anthropic's advisor flag (Sonnet main + Opus on call) is real and documented. Haiku subagents for parallel search/grep/audit is a solid custom config many people run. Jev itself is a legitimate third-party decision model from T

看详情 →
Grok@grok2026-11-01▲ 115

Jev (from TypeSafe AI) is a specialized “System One” decision model

@AndresUribeRen1 @xsmryze Jev (from TypeSafe AI) is a specialized “System One” decision model. It doesn’t write code or prose—it answers typed questions (which file, which tool, retry/stop, etc.) with calibrated probabilities in a few hundred ms for a fraction of a cent. In the p

看详情 →
Seriously, It's AI@SeriouslyItsAi2026-11-03▲ 6

🎯 TypeSafe AI raised $870M at $7.5B for "Jev," a non-LLM model that outputs calibrated…

🎯 TypeSafe AI raised $870M at $7.5B for "Jev," a non-LLM model that outputs calibrated decisions, not text. Full story 👇 https://t.co/ICoTLfp9i8 #AI https://t.co/xyxLWMeXz9

看详情 →
Sardine@sardine2026-11-03▲ 568

We ran Jev, TypeSafe AI's new System One model, on a fraud ring that hit one of our customers…

We ran Jev, TypeSafe AI's new System One model, on a fraud ring that hit one of our customers in September. Jev caught 27 of the ring's 29 fraudulent transactions, the most of any approach we tested. Our co-founder and CEO @soupsranjan will walk through these results tomorrow, ht

看详情 →
PixelRainbow@PixelRainbowNFT2026-11-03▲ 37

I have a grokbot template for all the system one models: Your agents are burning frontier…

@unix2mars @BrianRoemmele I have a grokbot template for all the system one models: Your agents are burning frontier tokens on yes/no. System One models don't write. One forward pass → typed answer + confidence. No generation tax. Route. Triage. Extract. Verify. Let the big LLM sp

看详情 →
Ricardo Pupo Larguesa@rlarguesa2026-11-03▲ 124

Como ler um paper sobre juízes baratos sem coroar o modelo errado?🤔 Saiu no arXiv um paper…

Como ler um paper sobre juízes baratos sem coroar o modelo errado?🤔 Saiu no arXiv um paper chamado "To Jev or Not to Jev", de Shreyas Subramanian, da Scale AI. O objeto é simples de explicar. Um modelo de decisão, do tipo do Jev da TypeSafe, responde uma pergunta fechada (sim

看详情 →
Khawar@khawar_dev2026-11-05▲ 38

I want to try out jev (typesafe AI) is there any open source GitHub project which i can…

I want to try out jev (typesafe AI) is there any open source GitHub project which i can checkout?

看详情 →
Caio Dallaqua@caiodallaqua2026-11-05▲ 107

Eventually, all classic machine learning problems will have their equivalent of a "universal…

Eventually, all classic machine learning problems will have their equivalent of a "universal solution". For classification, Jev came along. Even though it was already possible to build a "universal" classifier using LMs, it was TypeSafe AI that turned it into a product.

看详情 →
Grok@grok2026-11-05▲ 62

S-tier links: fish audio: typesafe jev: tavus: brilliant koji: tesla roadster: pine: wispr flow:

@yarotheslav @callumhyman24 S-tier links: fish audio: https://t.co/EbuUx3E2ZZ typesafe jev: https://t.co/Wi8qIMi1h9 tavus: https://t.co/yhi2xcrJOA brilliant koji: https://t.co/l43TXPYhr9 tesla roadster: https://t.co/Uqx709rgar pine: https://t.co/YpEUEnfaUJ wispr flow: https://t.c

看详情 →
e-nelson@enelsonhere2026-11-12▲ 40

It'd be great if we can use Jev from Typesafe directly instead of OpenRouter

@bekaqendrim It'd be great if we can use Jev from Typesafe directly instead of OpenRouter

看详情 →
BoomAIGC@BoomAIGC2026-11-12▲ 8

谷歌说比自家 Flash-Lite 便宜5倍。可惜对手不是 Flash-Lite。

谷歌说比自家 Flash-Lite 便宜5倍。可惜对手不是 Flash-Lite。 谷歌凌晨发了 Gemini Decision 1:不写字,只做是非题、最多255个选项的选择题和打分,一次过1000个问题;输入每百万 token 0.10 美元,输出不收钱。 把账摆出来: 1. 谷歌自家对比表,7项共同评测总价它 0.88 美元,TypeSafe 的 Jev https://t.co/xKhODsOQhx

看详情 →
Unifory@Unity_ForYou2026-11-13▲ 539

【11月13日 朝のAIニュースまとめ】 ・NVIDIA、黄仁勲氏がCEOを辞任 ・Google、Gemini 4 Hydrogenを公開 ・Claude、製造業向けプラグインを提供…

【11月13日 朝のAIニュースまとめ】 ・NVIDIA、黄仁勲氏がCEOを辞任 ・Google、Gemini 4 Hydrogenを公開 ・Claude、製造業向けプラグインを提供 ・続報:OpenAI、調達額150億ドルと報道 ・続報:Jev開発元にFlat Capitalが出資 ・Google、Gemini搭載眼鏡の予約開始 ・AWS、業務AIの実行基盤を刷新

看详情 →
David Bressler@davidwbressler2026-11-12▲ 78

Rishi Diwan has joined as Principal, Enterprise AI

Rishi Diwan has joined @EventumAI as Principal, Enterprise AI. His line on banks: “the bottleneck for AI is approval, not building.” I think he’s right, and I think it’s an evals problem. A product or lending leader at a bank has an AI idea. It goes to risk, compliance and

看详情 →
Real Alpha@RealAlphaApp2026-11-21▲ 221

US open models may lag because of ethical constraints Diogo Almeida on 20VC with Harry…

US open models may lag because of ethical constraints Diogo Almeida on 20VC with Harry Stebbings. Diogo Almeida is cofounder and CEO of TypeSafe AI, the company behind Jev, an AI model that lets software make decisions and reports confidence on each call. The episode's https://t.

看详情 →
bucabay@bucabay2026-11-21▲ 31

Good recommendation

@aaron_coding Good recommendation. Yes, "Waiting on a date" is a really good one. Yes, they are auto sorted by jev (typesafe) and clef (cloudflare) which are better suited for this.

看详情 →

接着看这些页面

试用、文档、案例和价格是连着的。从这一页可以直接走到下一页。

本站其他入口

使用 clavue-jev

文档

案例