Race-swapping in these movies is essentially the same as putting black faces on traditional European stories. It’s not fighting racism — it is racism https://t.co/ySveC8RZws

@spikeytigger @steipete Hit me up if you ever need support or have suggestions, or join our discord I’m there even more often https://t.co/SxWoPjZTVD

@n0bledeer @steipete https://t.co/pZsDRVmAqh

@PraveenInPublic @paulg Yeah, this is exactly why I built my lists: https://t.co/9eRY65x3IQ and my AI that finds the best out of them: https://t.co/8L5xphk0qQ and why I join so many audio spaces to share what's going on here in SF with others (and why I do so many long videos, like the one I just posted).

The UK has become a prison island. https://t.co/KQXA06XDWn

@ManuKumar @cyberpengk I'm automating all my inbound. It looks for leads and puts them into a kanban so I, and my team, can more easily track the status of people asking me to do things. Here are a set of use cases that my AI gave me: +++++ Here are 20 real documented use cases — drawn from the OpenClaw intelligence report (76 cases researched in February) and Hermes user reports from May 2026. Most work on both platforms. (Done by Levangie Labs' AI Agent that I built). DEVELOPER WORKFLOWS 1. Automated code review on every commit — Agent watches your repo, flags issues, messages you a plain-language summary. Greg Brockman's "run Codex on every commit" is exactly this pattern. 2. Error log monitoring — Agent watches production error logs, identifies patterns, proposes fixes directly to your phone before you've even noticed the problem. 3. PR summarization — Agent reads open pull requests and sends plain-language summaries to iMessage/WhatsApp so you can approve or reject without opening GitHub. 4. Local model routing — Agent routes simple tasks to cheap local models (Qwen3.6-27B on your Mac Mini) and escalates complex reasoning to cloud. Power users on Hermes report 10x cost savings vs API-only. BUSINESS OPERATIONS 5. Email triage and response drafts — Agent reads your inbox, categorizes by priority, drafts responses for high-value emails, sends drafts via iMessage for one-tap approval. 6. Meeting prep briefs — Agent researches attendees, pulls relevant docs, summarizes context, and delivers a brief 30 minutes before each calendar event. 7. CRM automation — Agent listens to call transcripts, auto-updates Salesforce/HubSpot records, flags follow-up actions, and pings you if a deal goes cold. 8. Competitive intelligence — Agent tracks competitor job postings, product announcements, pricing changes, and fires an alert when something material happens. CONTENT & MEDIA 9. Newsletter/briefing compilation — Agent aggregates sources you define, drafts the weekly briefing, sends for your edit. (This is essentially what https://t.co/dhsLUWXlkp does, run autonomously.) 10. Social media signal surfacing — Agent monitors specific accounts and keywords across X/LinkedIn, surfaces high-signal posts to your phone. Your own OpenClaw skill would do exactly this for your lists. 11. YouTube/podcast transcription pipeline — Agent transcribes new episodes from creators you follow, extracts key insights, adds to your searchable knowledge base. PERSONAL PRODUCTIVITY 12. Smart calendar management — Agent schedules meetings based on your preferences, handles conflicts, and negotiates timing on your behalf via email — without you touching the calendar. 13. Research synthesis on demand — Ask a question via iMessage, agent searches web + papers + your personal documents, delivers a synthesized answer with sources. 14. Personal memory layer — Agent captures everything you read, forward, or annotate and builds a searchable memory. Ask "what did I save about biosynthetic pathway design last month?" and get an answer. FINANCE 15. Portfolio and market alerts — Agent monitors positions, flags significant moves with context (not just price — why it moved), keeps you informed without you watching Bloomberg all day. 16. Invoice and receipt processing — Agent reads incoming invoices, extracts data, flags anomalies vs budget, queues approvals in your accounting system. HOME & SPECIALIZED 17. Smart home orchestration — Agent controls lights, temperature, appliances based on time, weather, and routine. More sophisticated than Alexa because it reasons about context, not just commands. 18. Health tracking synthesis — Agent integrates Apple Watch, Oura, and sleep data, suggests when to train hard vs rest, flags patterns you'd miss manually. 19. Real estate deal monitoring — Agent watches Zillow/Redfin for listings matching your criteria, runs comp analysis, alerts with a recommendation before the listing gets crowded. 20. Contract and legal document review — Agent reads PDFs, flags unusual clauses, asks targeted clarifying questions, and drafts a plain-language summary for your lawyer to confirm. The biggest shift Hermes users report vs OpenClaw: fewer agents running better. OpenClaw's 400-agent architecture was powerful but fragile. Hermes users are building 3-5 deep agents that actually learn their workflows vs dozens of shallow ones that need constant maintenance.

@steipete Also there’s this wolfbench too by weights and biases. But do any of these benchmarks matter? I’d argue generally no! It’s the user experience and the community experience that matters, and you have failed. Thats part why your token throughput has done nothing but collapse, starting the day Hermes was released, and in just 3 days since surpassing you, we’ve nearly 2.5x’ed your token volume. Real users using it. They chose.

American science is at extraordinary risk. NIH has awarded less than half as many grants as it has compared to the past five fiscal years averaged together. 'I thought we were at rock bottom', the official said. 'We are below rock bottom now.'" https://t.co/HkGjVoIsGa

My AI at https://t.co/kiuZ7QXLzb told me to read this paper before watching Google's IO next week. My AI is always giving me things to do, isn't yours?

@steipete I’m sorry that you’re this desperate that you will take such an unscientific benchmark instead of any established one. Also qwen local is one of the most random length models there is with all its looping. And we smoke you all on quality benchmarks on every open model. Here’s wildclawbench by internlm, same speed on open models, much better results.

An Econ PhD student at the 20th ranked program who is working on stuff they are passionate about will have a better job market than one at MIT who's been doing nothing but phd-app-maxxing since undergrad. People get confused by this because they don't observe *how* successful people came about their insane knowledge bases. It wasn't by relentlessly grinding away at stuff because they had to. They look at Scott Kominers and say "if i grind and learn as much math as he did, i will be successful." You can't! *You* can't learn as much math as Kominers because he gets energized by configuration results for type ii lattices. You will burn out if you try to do it this way. You cannot, through grind alone, learn more about the economics of cities than Glaeser, or about how to maximize a value function than Acemoglu. Research careers are long. Most people give up and stop working on research (graph is share of elite PhD graduates with at least one publication in year X after graduation). If you're starting a PhD, you're presumably doing it to have a successful 40-year research career. The number one factor in whether that happens is not which program you get into, it's whether you find a research angle that energizes you enough to push through the endless barriers an academic career throws in your path. This is why a lot of the received wisdom around PhD applications is wrong. If you're 100% consumed by the predoc rat race already, it's going to be a long, hard road ahead. Obv you still have to do admissions, you should study a lot for the GRE, sigh it seems like taking real analysis is probably worth it. But spending time on the things that energize you about economics is a no-brainer, whether it's policy, or blogging, or whatever, you gotta do the things that light your fire and make you want to be on this road.

Woke is a death cult that executes common sense and devours anyone who tells the truth It spreads insanity like a plague and turns mental illness into the new standard Worst of all, it has turned pure evil into a religion and demands we worship it https://t.co/ipoiP2w0HH

Constructing is far nobler than deconstructing. To the extent that the West is faltering, it is only becuase takers have gained more power than makers Building is moral. Back to building. Never fall into the poisoned mental frame of those who cannot want and imagine a positive future, and therefore have no willpower to build it.

+2 more

AI animation is getting ridiculously fun. Create retro arcade sequences with pixel graphics, vintage game transitions, giant bosses and rainbow power-up modes. The ultimate retro pixel animation PROMPT ↓ https://t.co/gtJVfrDzZT

The Woke Mind Virus in Academia https://t.co/tAEW2eE0P1

Grok Build has three commands for managing memory across sessions: /memory, /flush, and /dream. They're experimental but worth looking at if you've ever been frustrated with how agents forget everything between conversations. /memory opens a window into what Grok has saved. There are three layers: global memory, workspace-specific memory, and per-session summaries. You can read what's there, edit it, or delete things you don't want kept. /flush is for when you've had a useful session and want it saved before context compaction kicks in. It writes a summary of the current conversation into the memory store, capturing decisions, debugging paths, project conventions, and anything else worth keeping. /dream runs in the background over your old session logs and memory fragments. It deduplicates overlapping notes, merges related fragments, and consolidates everything into cleaner topics, so the store doesn't grow into a pile of half redundant snippets over time. Most agent memory I've looked at just shoves the chat history into RAG. That works for about a week before the store gets noisy and starts hurting sessions. Grok Build treats capture and consolidation as separate commands, with the user able to inspect what's saved. The editable part is important. If your agent saves something wrong, or if your conventions change, you need to be able to go find that memory and remove it. Otherwise the agent keeps applying outdated context with full confidence and you spend cycles undoing its mistakes. Managing context for long-running agents is going to need real memory primitives. Write, search, prune, and consolidate, all as first class operations. Grok shipping these three commands is the first time I've seen a consumer product treat memory as its own layer.

MIT announces “the number of grad students will be 20 percent less than it was in 2024 — about 500 fewer students” https://t.co/Hk0qzQ53eL

@highbrow_nobrow @sandy7beach They can only win by cheating. Deplorable! https://t.co/suK9JifIdw

@highbrow_nobrow @sandy7beach They can only win by cheating. Deplorable! https://t.co/suK9JifIdw

We want our country back . Millions are in attendance https://t.co/29EKkv5d3e

A decade ago, Hollywood decided it would no longer honor films based purely on merit but also on the diversity of the cast and production team. In 2020, they formally tied Best Picture to diversity. The Academy expanded its membership by roughly 40% in a decade through an explicit push to diversify its ranks. The Oscars will proudly tell you that last year’s invitees were: >41% women >45% people of color >55% from overseas For 88 years, the Academy selected members and honored achievement based on merit. For the last 10 years, invitees went from 112 per year to over 900 to meet diversity targets. This devalues and dishonors great American films and creates perverse incentives that lead studios to inflate diversity numbers in hopes of recognition by their woke peers. Meritocracy must be restored to film.

xAI got some seriously best designers now The design quality has really leveled up and it’s pretty sick https://t.co/S3xVs3rnGY

When in New York! https://t.co/FNrmkOlLm5

After the Reform Party HUMILIATED Labour in elections, Keir Starmer smeared the ‘Unite the Kingdom’ rally as “extremist” and banned Conservatives from entering the country. The people of England’s response?: Tens of thousands FLOODING the streets 🇬🇧🏴󠁧󠁢󠁥󠁮󠁧󠁿 https://t.co/lNTxhLeT7q

Everyone has seen the @waitbutwhy cartoon of AI capability growth with a "you are here" indicator just before the exponential really starts, but the independent assessments of both METR and the UK's AISA do seem to show that we are past that point now (until we hit a slowdown?) https://t.co/vxYc6GSS5d

+1 more

NVIDIA just released a paper review dataset on Hugging Face APRES, Agents4Science, and Sakana v2 subsets covering human and AI-authored papers with real review decisions. https://t.co/louiNyRGC5

AI will make translation invisible. But language was never just about translating words. Language is trust. Culture. Humor. Negotiation. Influence. The people who understand this will still have a major advantage in business and life. AI removes friction. Human connection remains the moat. https://t.co/KjuTAKxG6x @ConversationEDU @ConversationUS

New article: a visual tour of recent LLM architecture advances, from Gemma 4 to DeepSeek V4. I focus on long-context efficiency tweaks like KV sharing, per-layer embeddings, layer-wise attention budgets, compressed attention, and mHC. Link: https://t.co/KO81y3kTH7 https://t.co/wTx51QpQu4

Your Grok Subscriptions work inside of Hermes! https://t.co/TGbJMDzTMI

Your Grok Subscriptions work inside of Hermes! https://t.co/TGbJMDzTMI