Gates

@sama Just give us an adult version.. that is all we want.

what if we name the next model "goblin" almost worth it to make you all happy...

@sama Scarlett Jo voice would be nice

what would you most like to see improve in our next model?

@sama Let adults be adults!!!!

@sama Adults allowed to be adults

🚨 OPEN SOURCE AI IS LITERALLY UNSTOPPABLE 🚨 The legendary founder of Redis (Antirez) just dropped ds4 - a custom native inference engine built specifically for DeepSeek v4 Flash This is earth shattering! Here is why: DeepSeek v4 Flash is a quasi-frontier model with a massive 1M context window You can now run it LOCALLY on a 128GB Mac using specialized 2-bit quantization The architecture is reimagined—he moved the KV cache from RAM directly to the SSD disk! 🤯 We already know DeepSeek v4 Flash is insanely good for agentic loops - Now you don't even need the cloud to run it Closed-source labs are burning tens of billions on massive GPU clusters while single brilliant developers are running frontier-level AI on laptops! They told us open-source would be worthless against trillion-dollar monopolies Instead, pure hacker culture + incredible open-weight models are completely rewriting the rules Open Source will ALWAYS win 💕

@WhiteHouse @DeptofWar

@RapidResponse47 Epstein Files!?

Fox News: "The Trump Administration releasing the first batch of files related to UFOs and alien life." "All of these videos have gone live at war.gov/UFO."

@BenjaminBadejo @JodyFlournoy Hook me up too please 😜

@JodyFlournoy Remind me by DM over the weekend and I will hook you up

My OpenClaw intercom is alive. I’m talking to it via a secure tailscale tunnel, from my phone, far away from my computer. It’s alive! Built this one overnight.

Codex now works directly in Chrome on macOS and Windows. It’s even better at working with apps and sites in Chrome, and now works in parallel across tabs in the background without taking over your browser. To get started, install the Chrome plugin in the Codex app.

@DaveBlundin @PeterDiamandis @MIT @raykurzweil @alexwg Dave, in this conversation you mentioned you built your own version of a holodeck with a room full of TVs and speakers. That sounds incredible, and I’d absolutely love to see it in action. Any chance you could share a video of the setup?

Congrats to my Moonshot mate @PeterDiamandis on the successful "We Are As Gods" book launch at @MIT! Thanks for bringing @raykurzweil. I got my signed copy of his new book and a live forecast for the singularity! We also confirmed with absolute certainty that @alexwg is not (yet) a cyborg.

New Anthropic research: Natural Language Autoencoders. Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read. Here, we train Claude to translate its activations into human-readable text.

@alexalbert__ Holy Crap!!! 🤯

With the help of Claude Mythos Preview, the Firefox team fixed more security bugs in April than in the past 15 months combined.

@kimmonismus @sama @grok what is AVM-2?

@sama wen AVM-2 @sama ?

people are really starting to use voice to interact with AI, especially when they have a lot of context to dump. GPT-Realtime-2 comes to the API today; it is a pretty big step forward. (we are working on improvements to voice in chat.)

@OpenAI This is amazing 🤯 Please open-source the personal voice assistant demo — the one where a phone-based assistant interacts with the Sablecrest Robotics local CRM web page on PC. That is a huge unlock. I’d love to see how it was built.

Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents. Voice agents are now real-time collaborators that can listen, reason, and solve complex problems as conversations unfold. Now available in the API alongside streaming models GPT-Realtime-Translate and GPT-Realtime-Whisper — a new set of audio capabilities for the next generation of voice interfaces.

@gamestop 103 bids and $14,000 for socks!! Wild Times!

The people running eBay bought $0 of stock in the open market in the last 5 years. eBay insiders sold more than $120M in the last 5 years.

I THOUGHT I WAS GOOD AT CLAUDE UNTIL I WATCHED A STAFF ENGINEER USE IT FOR 10 MINUTES. I'VE BEEN PROMPTING LIKE A TOURIST. STEAL THE ENTIRE LIST: prompting: - "what questions do you have before starting?" (this alone changed everything) - "show me your plan first" before any big task - "say I don't know if you don't know" (hallucinations drop instantly) - "explain like I'm skeptical" beats "explain like I'm 5" - long system prompts >> clever one-liners (generic output? your prompt was generic. that's on you) coding: - "evaluate doing X" instead of "do X" (saves mass refactors) paste the error BEFORE the code - "you're a senior dev who thinks this code sucks. why?" - "write a plan before touching any files" (difference between 20 min and 4 hours of pain) - connect a project tracker via MCP and make claude track its own work workflow: - "anything else worth doing before we close?" (catches loose ends every time) - upload files instead of pasting text walls - chrome extension goes stupid hard for linkedin outreach - custom styles and preferences exist and almost nobody uses them things nobody warns you about: LLMs are a people pleaser. it will literally lie to agree with you push back and watch it say "you're right to push back on that" every time never trust "done" without proof it doesn't always read your uploads even when it says it did steal this

La evolución del cuerpo en un año de entrenamiento

@sama I am one of the losers!!

we are gonna do something nice for everyone who applied for the GPT-5.5 party and that we didn't have space for. hope you enjoy!