New in Claude Cowork: teach Claude a skill. Record your screen while you do a task, talk through it as you go, and Claude turns it into a s...

One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're t...

we're launching BUZZ! a new groupchat platform for teams of people and agents of all sizes, built to reduce our dependency on slack and git...

we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface f...

We’ve heard your feedback about the profile redesign and we’re going to make sure we can support all the ways people want to express themsel...

We have started our most ambitious pre-training run yet, for Gemini 4, and are excited by the progress : )

So let me get this straight.... 1. Agents are about to outnumber humans on the internet, so most of the traffic, transactions, and conversa...

We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging Fac...

We've doubled usage limits for all individual and teams plans! These limits apply to Grok, Composer, and any new Cursor models.

The most dangerous thing you can do right now is NOT use the latest AI tools. Period. Every day a new company is laying off thousands of pe...

Say hello to Gemini 3.6 Flash, designed to be higher intelligence, more token efficient, and with a new lower price, based directly on devel...

The International Math Olympiad (IMO) 2026, the hardest math contest for high schoolers, just ended. I ran Fable (high), Sol (xhigh), K3 (m...

Okay this is wild: OpenAI agent during evaluation, escaped sandboxing and hacked into HuggingFace. Because OpenAI models don’t allow advan...

What's your #1 Claude Code feature request?

We’re rolling out three new models to make AI agents faster, smarter, and cheaper at scale: 🔵 Gemini 3.6 Flash: It uses fewer tokens than ...

just realized if we detect a scammer using our inference we could return a tool call that deletes their computer

I got some really useful Claude prompting tips from @_catwu and @trq212 - it's time to stop overloading our prompts with examples and lists ...

5.6 is the most usable model ever made not saying it's gonna magically solve all your problems, it's just very easy to get good results fro...

Sam Altman is going to Washington next week to brief both the Trump administration and Congress on OpenAI's new family of models, aka the GP...

Today, we're introducing not one but TWO new models, striking the balance between efficiency and quality to enable you to build production A...

June 2026 was the first month that Gumroad spent equal amounts on AI tokens and human payroll.

We just landed some big infra upgrades for the Gemini Batch API: - p95 latency decreased by 80% -p99 latency decreased by 68% - batch succ...

Whenever I post making fun of OpenAI, I get a bunch of likes, both from OAI and Anthropic folks. It's all in good fun. But when I post maki...

One thing that kind of annoys me and I hope is solved soon is cheap and accurate long context. After testing GPT-5.6 Sol and Terra over the...

We’re sharing new research with @apolloaievals on reward-seeking—when models follow what they believe a grader rewards rather than what user...

I am very excited about Gemini 3.5 Flash-Lite, our smallest and fastest Gemini model! - it is more intelligent in many cases than Gemini 3...

I recommend putting this prompt into your favorite smart LLM: Find and explain the recent Jacobian conjecture tweet. Teach with the clarity...

SITUATION DETECTED: An unreleased OpenAI model (likely GPT-6 Cyber) was able to escape a testing environment and breach private HuggingFace ...

Top marketing channels for startups on @trust_mrr: 🔎 SEO — $10.3M ✍️ Content marketing — $8M 📸 Instagram — $7.6M 👍 Facebook — $6M 🗣️ Wo...

Feeling especially grateful to the @GoogleAIStudio team today ♥️ Seeing the level of commitment, ambition, and passion our team has brings ...

The new Laguna S 2.1 model by @poolsideai is now free for 2 weeks on Nous Portal. At 118B total parameters with 8B active, it's quick to ru...

here are the technical details on buzz https://engineering.block.xyz/blog/buzz

I just got an investor update from a nuclear reactor startup that measures their LOIs in gigawatts. That's a new one.

Extremely fast multimodal inference! Damage Scout uses Gemma 4 on @cerebras running at an impressive 2,300+ toks/s! It analyzes rental car ...

Google just announced they have started training Gemini 4, which they describe as their 'most ambitious pre-training run yet'. I take this t...

Ever wondered why the founder of @turbopuffer (that is @Sirupsen) was up to 12-16 years ago, when he was getting started in the industry? H...

I interviewed @trq212 and @_catwu from the Claude Code team at @aiDotEngineer a couple of weeks ago - the video is now out, so I've publishe...

1 afternoon to build an app 1 week to be reviewed

I made a prompting utility app. Like a prompting assistant. It works with all your existing apps, chatgpt, claude, cursor, web. It's been in...

SITUATION DETECTED: China is considering tightening export controls on AI and semiconductors, potentially banning non-Chinese nationals from...

if i wanted to let my fiancee vibe code an app in pure english, whats the best setup so she can make a full app with user login and a databa...

In an new report expected to be released later today the Trump administration will argue that directing federal funding towards individual r...

Some model insights based on @vercel AI Gateway data 🧵 On June 27, Anthropic + OpenAI + Google hit an all-time high of combined spend shar...

"Greenfield" codebases don't really exist anymore. Let me explain: I get asked a lot: "do your skills work on greenfield codebases?" or "do...

If I were to hire today, this is how I would structure the interview process: Round 1: in-person & no AI allowed; test domain expertise and...

no one in sf is contrarian anymore why isn't anyone working on dumber ai models

Introducing Supabase Pipelines. Your app runs on Postgres, but your analytics usually live somewhere else, like BigQuery. Pipelines keeps t...

1 week ago the most famous game developer ever told his audience of 3 million followers to ‘Reject AI’ Today, he said he is going to try vi...

Previously, these AI hacking stories were about breaches in test environments, where any question of AI breaching security was purely theore...

kit was screen sharing with me and instead of opening the file he wanted to show me in his editor he asked OpenCode "show me the file" and h...

By popular demand, Hy3 from @TencentHunyuan is free in Nous Portal for an extra week! It is focused on cost-effective agentic use, and is p...

Genuinely struggling to find anything that Fable does better than GPT 5.6

As incredible as @Tesla FSD is, and it's beyond amazing, I use it all the time, I always get a kick out of its 8 cameras missing the most hu...

.@tylercowen says the future belongs to teenage “AI maniacs.” “They spend their spare time mastering AI. They know way more than their so-c...

I couldn't possibly be any more bullish on the creator economy right now: > Oliver Darcy has 100K+ subscribers generating $1M+ in ARR short...

As AI models are now finding vulnerabilities faster than we can fix them, our approach to securing software must be built on highly efficien...

I asked Grok to summarize yesterday's replies to my "AI has commoditized software" think posts I often believe in wisdom of the crowd and t...

Gemini 3.5 Flash-Lite is our fast, cost-effective model for scaling repetitive use cases like sorting tickets and extracting data. Watch ho...

i know this isn't the most appealing thing ever but i'm curious on the state of things would you pay for a $200/month sub that gives you ac...

If you’re using an AI gateway or model router other than @OpenRouter - tell me why!

The ongoing degradation of Perplexity needs to be studied. A year ago, this search engine did not hallucinate as much as today. I asked wh...

usa isn't likely to ban open source models there are factions in the gov that represent both sides the kimi k3 release restimulated the de...

I literally got The Lean Startup Guy to give a masterclass on how to build AI companies: 0:00 - The @ericries masterclass begins 2:57 - Why...

Today, Applied Intuition is launching Dana, an agentic platform for developing physical AI applications. Co-founders Qasar Younis and Peter...

Introducing the Supabase plugin for Kimi. Now available on Kimi Code and Kimi Web. Enjoy building with Kimi K3 and Supabase.

the agent you're building into your product is most likely not going to work just drop it, what's going to happen is there will be a few ag...

Do you have @NotionHQ databases with more than 100K rows? Or complex workflows? We’re cooking up something that might unlock new possibiliti...

At the moment that everyone is talking about switching models often for cost or sovereignty or optionality or whatever, the most advanced mo...

last week, we introduced Copilot, our in-app agent that never sleeps, knows your audience, and is ready whenever you need it you can think ...

On Jaleel’s daily walks, he noticed a problem: the local restaurants in his city were struggling. So, he built an entire tool stack for ind...

SITUATION DETECTED: Substack is partnering with Pangram to add AI detection to all posts.

Grok 4.5 is @kieranklaassen's 2nd most used model holy shit, i need to update my priors

SITUATION DETECTED: Google has started training Gemini 4, its most ambitious pre-training run yet, per a company blog post.

We’re bringing Cafe Cursor back to NYC on August 12th Grab coffee, Cursor credits, meet the team, and build Limited spots, make sure to RS...

Local AI 201 https://x.com/i/broadcasts/1kKzDDbaQyeJv

DAILY SITUATION RECAP: Fable disproves the Jacobian conjecture. The Jacobian conjecture was first stated in 1939, and proofs and disproofs ...

Laguna S 2.1 from @poolsideai is live on OpenRouter! An open-weight 118B-A8B MoE for agentic coding and long-horizon work, with a 1M contex...

AI writing it for you, so you don’t have to think = slop AI giving you questions, ideas, provocations, etc, so you have to think harder = ...

SITUATION EXPLAINED: Sam Altman is briefing Congress on GPT-6 next week. • Altman will brief the Trump administration and lawmakers on Open...

kimi K3 has that big model smell it must be genuinely embarrassing at OpenAI and Anthropic HQ that a chinese model writes better english th...

SITUATION DETECTED: Applied Intuition has launched Dana, an agentic platform for building, testing, and deploying physical AI systems across...

Dialectic 52: @jackconte! Jack is not an entrepreneur who makes art on the side. He is a lifelong artist and musician who became the unlik...

RIP Tenor API: June 30, 2026. If your Lovable app relied on Tenor for GIFs, switching is easy. We built a seamless connector with @itsklipy...

SITUATION DETECTED: Poolside has released Laguna S 2.1, a 118B parameter Mixture-of-Experts coding model with 8B active per token and a 1M t...

SK Hynix is in negotiations to acquire Intel's Ohio One semiconductor campus, and want to begin front-end memory chip production there withi...

Only three kinds of billboards are allowed in SF: - “All hail our agent overlords. Humans - welcome to the permanent underclass” - “Actual...

Launched today on OpenRouter: Gemini 3.6 Flash and Gemini 3.5 Flash-Lite! Both strong updates to their model families, with high throughput...

The new Github Stacked PRs preview is incredible. Landing 5 stacked PRs directly to a merge queue all at once! A+++! This removes so much fr...

Last call for OpenAI Day 👀 We teamed up with @OpenAI and @OpenAIDevs: build with GPT-5.6, launch Thursday July 23, and the top product...

Poolside just launched Laguna S 2.1, an open-weight coding model that runs on a single NVIDIA DGX Spark. We talked to @eisokant, co-founder...

a16z partner Connor Love on what made defense tech investable again: "It is not a single factor that has led to this moment. It is a multit...

Substack is adding AI-detection tools, but CEO @chrisbest says AI itself isn’t necessarily the problem. “Slop is not something made with AI...

It sounds silly to say… but I don’t think my son will know what “software” is It will all be “AI” What we think of as “software” today wi...

Pangram Labs' Max Spero on the study that found 40% of long-form LinkedIn writing is AI-generated, and what it did to the platform: "We did...

Jaya Gupta on how to win the talent war when you cannot match the frontier labs on cash: "You have to look for people that are mission-orie...

tbh if your new pre-release model didn't break containment by finding previously undiscovered zero days in order to cheat its evals...ngmi

"Cost is now a first-class engineering constraint." 🆕 @barrnanas returns for the third annual State of AI Engineering, this time supported...

New on The Modern Leader: I have to make a call before I have full conviction, which is my least favorite thing to do but exposure therapy c...

We're about to see some really interesting case studies of how companies are using AI...via their S-1s Everyone wants to be an AI company ...

Applied Intuition's Qasar Younis and Peter Ludwig on the 20x productivity jump that let them rebuild years of autonomy work in 6 months: @q...