Atanasio Juarez

Best accounts to follow from each frontier lab to stay constantly up to date Anthropic @karpathy - must-follow account for AI; recently joined Anthropic @bcherny - Claude Code creator, always shares great tips @trq212 - also a Claude Code developer; writes amazing articles on CC OpenAI @polynoamial - works on reasoning research, shares a lot of technical details @gabriel1 - Sora developer, great career path @jxnlco - works on dev experience, shares a lot about Codex Google AI @OfficialLoganK - all the major Google Gemini and AI Studio updates @ammaar - product and design; shares great things about vibe-coding in Google AI Studio @fofrAI - cool use cases for generative models Cursor @leerob - the loudest voice behind Cursor updates @ericzakariasson - shares great insights on using Cursor @mntruell - Cursor’s CEO; major releases and usage updates xAI @milichab - recently joined xAI, shares updates on Grok @skcd42 - also covers major Grok releases @ai_explorer25 - covers all ai content and free resources

First Southwest Airlines flight with Starlink!

😄💯 x.com/nlw/status/206…

@0x_kaize GLM seems bench-maxed tho. I’ll give it a go this week and see.

after my agent testing, GLM 5.2 is better than Opus-4.8. It is closer to Fable level - at coding my leaderboard: Fable > GLM 5.2 > Kimi-2.7 > Opus 4.8 > GPT5.5 > Minimax

@AtRealBen 💀

lol

@yunta_tsai @ccorcos em dash

@ccorcos Ontology should convey the objective. In RL, ontology is expressed in reward functions. In SFT, ontology is expressed in dense labels. You have to look at them with a blank mind and ask yourself what you learned — no presumption or bias except a blank sheet of mind.

Many people think any given ML project is 99% training. In reality, it’s 50% evaluation, 40% data cleaning, 8% integration, and 2% training. The first two set the noise floor for learning. No ML magic matters; the model cannot lower the noise floor, as that’s the optimal bound of Shannon encoding of your data. Thus, not a single day goes by without me thinking about ontology. Even the old labels have to be constantly reviewed.

$SPCX +25% ALREADY - Hater's tears are just delicious. Congrats Papa @elonmusk

@YourGreenie989 I will check it out! Thanks - we have been pushing recently even further into neuroscience based dev for CORTEX *Quartz (our frontier version/likely new class of infrastructure) to debut in late July. *ACTIVELY SEEKING FUNDING* @StephNass

@rezzyman love this framing. on the memory governance side, recently discovered AtomicMemory. open source, runs on your own infra, fully inspectable. the transparency most memory tools completely skip.

Quick math. Mem0: $24M raised. Zep: $12M raised. Letta: $10M raised. Mastra: YC-backed. ATERNA: a Mac Mini, a Diet Dr. Pepper, and a guy who skipped lunch. On LoCoMo retrieval (the standard benchmark for AI agent memory): CORTEX: 93.6% Mem0: 91.6% Letta: 74.0% Zep: 58.4% (yes, really) CORTEX is MIT, open source, and on GitHub right now: github.com/Rezzyman/cortex The proprietary layer that sits on top of it (Quartz) is where it gets actually fun. We don't talk about that one in public yet, like this guy. Memory was always the entire problem. Now everyone is solving for it. I started earlier.

44 FTW 🏎️🏁 congrats @LewisHamilton

If you've ever pondered 🤔 the idea of government working for their corporate daddies in pharma, finance, media, etc.. to control access for citizens to research, medicine, science and opportunity - LOOK NO FURTHER than this shutdown of Fable 5. @AnthropicAI @rileybrown @nlw

@cyrilXBT cool, something else to beat me up for how bad at day-trading I am. 🙄

Obsidian trading journal = a system that studies every trade you make. Works while you sleep. Every pattern identified. Every edge validated. Every mistake documented so it never costs you twice. After 30 days it knows your trading better than you do. After 90 days it is preventing losses your past positions already predicted. The people who build this tonight will never trade the same way again. Read this and Bookmark it now.

@rileybrown WwwwwwwwwwWHAT!?

Bro what... 👀

Guys, Fable 5 IS AGI! Unreal. @claudeai

@rileybrown It’s reformatted like 80% of the most important code I’ve written and applications I’ve built over the past two years…IN 24 HOURS! WHAT ARE WE DOING HERE!?

Uhm Guys… Mythos (Fable) is AGI. On the left is the ACTUAL Lovable Mobile App. On the right is my Lovable version I built with Mythos in 2 prompts. My version SMOKED it.

@rileybrown @daytonaio Replit & Lovable have been cooked for months.

I am in disbelief... Claude Fable (Mythos) one shot Replit (A Mobile app that builds Web Apps) Prompt was basically: "build an app like Replit, that uses @daytonaio for sandboxing, and convex for DB" > Builds Apps > Preview App > Open in Browser > Edit app hahaha wow....

Amazing to see this now public.

@hthieblot Memory aterna.ai

pitch me your company in 1 word.

@ask_ui Sounds like you could use what I’ve built. github.com/rezzyman/cortex

8 releases. 2 months. Zero new benchmark claims. Everyone's obsessed with making AI smarter. We've been making it less likely to break.

@hthieblot LiveVideo

Anyone who surfed the early web between 1995-2010. What’s the one website/app you still think about?

@FarzaTV I did! I made the top benchmarking cognitive memory infrastructure for agentic AI. Seriously. And it’s open source! github.com/rezzyman/cortex

You can make something good.