@kshitijk4poor @AmeliaDacine Pushed a fix that attempts to solve it. Please try update: https://t.co/B9B12QgPm2

@Teknium Let's see where this friendship leads to. https://t.co/3vtvF0dt48

@Teknium Let's see where this friendship leads to. https://t.co/3vtvF0dt48

@Teknium @dingyi Last time you had a Hackathon I started editing Hermes and haven't stopped.. lol https://t.co/VQF3jjvqdG

@Teknium @dingyi Last time you had a Hackathon I started editing Hermes and haven't stopped.. lol https://t.co/VQF3jjvqdG

Steve Jobs, on why design matters more than raw capability: "The telegraph was an amazing breakthrough... some people talked about putting a telegraph on every desk... but it wouldn't have worked... you had to learn Morse code. It takes about 40 hours to learn Morse code." The telegraph was objectively more advanced for its time. Yet the telephone crushed it. Not because of superior technology. Because the phone required zero learning. "People already knew how to use it. Because you knew how to talk. You don't have to teach people how to speak." This is the insight most engineers miss: Raw capability doesn't win markets. Usability does. The telegraph demanded you become a specialist. The phone met you where you already were. This is why Steve was obsessed with making the Macintosh intuitive when competitors were shipping command-line DOS machines. The Mac didn't need a manual. It needed a mouse and a screen that felt like talking, not coding. You could put the most powerful machine on every desk. But if it requires 40 hours of training to use it, you've already lost. The winner isn't the breakthrough that's most technically impressive. It's the one that requires the least effort to understand.

New paper out in Proceedings of the Royal Society B: we apply linguistic tools to sperm whale vowels. The result: sperm whale vowels do not just look like human vowels. They also behave like them. We found several parallels. Like in Latin, whales have short and long vowels. Like in Slovenian, some vowels prefer particular tones. Like in human language, there’s a lot of coarticulation (a process when you say “tense” but the word sounds like “tents”). Observing vowels in whales is a matter of timing. Our vowels are fast, whale vowels are slow. Beats become pitch if they’re fast enough. If you slow down human vowels, they start sounding like whale clicks. Applying linguistic tools to whales shows us that we’re much more similar to these wonderful ocean creatures than we previously believed and that their language is much more complex and structured. @projectCETI @UCBerkeley

In a world ravaged by war, talk of the end of NATO, and an increased emphasis on financial sovereignty, AI sovereignty is growing in importance by the minute. McKinsey’s latest report on sovereign AI makes a compelling case for nations to build independent AI capabilities to ensure strategic resilience, economic impact, and societal benefits in an increasingly divided world. I give McKinsey credit for being spot-on and timely. Anyone working in AI should pay close attention to sovereignty implications in systems and solutions. AI sovereignty is also big business, and McKinsey estimates that 30 to 40% of AI spending could be influenced by sovereignty requirements a market of some $500 bn to $600 bn globally by 2030! While sovereign AI is often seen as a “defensive play” protecting citizens, data, and tech stack, McKinsey puts a new spin and says it should be considered going on the “offensive.” The US banning AI chip sales to China, China banning US LLM models, or the EU’s AI Act, should be seen as going on the offense for growth and innovation. The goal of sovereign AI ecosystems is to capture more of the economic value created by AI locally, rather than seeing it flow to foreign providers like the US’s Silicon Valley. Once again, there is big money involved. Nations going on the offensive with sovereign AI can capture a bigger piece of the $2.6T–$4.4T annually that AI contributes to global GDP. Sovereign AI isn’t just about protecting citizens and data from rogue foreign AIs; it’s all about getting a piece of the AI pie. Expect more nations to go on the offensive to bring AI’s money trail home. #fintech #tech #finserv #AI @BetaMoroney @efipm @BrettKing @spirosmargaris @jasuja @enricomolinari @mikeflache https://t.co/YM0i42elmo

Warren Buffett: "Most of the time, the Fed is not that important. Occasionally, it's everything." Buffett explained: "The Fed is of enormous importance during a panic." When the markets experience a crises, the Fed is the only person who can "walk out like a sheriff in the street and say that this isn't going any further." But in normal markets, Buffett said "people tend to hang on to [the Fed's] every word. But we don't pay any attention to it." Rather than forecast interest rates, he focuses on business fundamentals. "If the Fed Chairman were to whisper to me what his monetary policy was going to be over the next two years, it wouldn't change one thing I do."

Now in research preview: routines in Claude Code. Configure a routine once (a prompt, a repo, and your connectors), and it can run on a schedule, from an API call, or in response to an event. Routines run on our web infrastructure, so you don't have to keep your laptop open. https://t.co/m2XJWYqkf8

Jeff Bezos once asked Warren Buffett about Buffett’s investing strategy Bezos: “Why don’t more people copy your investment strategy? It’s not that difficult to understand in principle.” Buffett replied: “Oh Jeff, that’s easy. My approach is a get rich slowly scheme...” “The stock market is a device for transferring money from the impatient to the patient.”

The Microsoft split in 1983 that changed everything. https://t.co/Mqd4TjBCw7

The Microsoft split in 1983 that changed everything. https://t.co/Mqd4TjBCw7

Professor Jeffrey Sachs, makes the case that Trump is not only insane, but also a dangerously dumb human being. 😳👇 https://t.co/DYUyxA4Xq3

🚨 In 2 weeks, a final decision on amendments to the EU AI Act and the GDPR will be made. What is at stake is nothing other than the future of Europe. Many don't know, but the stream of events leading to this moment began much earlier, with the publication of the Draghi report on European competitiveness in September 2024. In his long report, Mario Draghi diagnosed various areas in which European competitiveness was lagging behind and suggested that one of the reasons was overregulation and the excessive number of laws governing the digital space. Laws such as the GDPR and the AI Act were to blame. Before I continue, here is something many overlook: the Draghi report was finalized in September 2024, while the AI Act was officially enacted one month earlier. The AI Act had barely been enacted, and it was already considered 'wrong,' excessive, and to blame for Europe's less-than-ideal (to be light) position in the AI race. From that moment on, the European discourse on the protection of fundamental rights was never the same. Its narrative shifted, and after that, the new dogma was that the path to innovation would be to "remove the red tape," and "apply the AI Act in a business-friendly way." (whatever that means from a legal perspective). The AI Action Summit last year made this new narrative loud and clear to the public, as EU officials abandoned fundamental rights-focused statements. Last year, the narrative shift was legally materialized. The EU published the Digital Omnibus with proposed amendments to some of its most important laws regulating data protection and AI: the GDPR and the AI Act. Strangely, the main justification for the AI Act's amendments was the designation delays by EU member states and the work delays by EU standardization organizations. If these were the real reasons, wouldn't it be more coherent to pressure them to move faster, hire more people, increase the budget, or help address the bureaucratic obstacles...? Does the EU need to amend some of the AI Act’s core obligations because EU bodies are delayed? I didn't buy it. Given the context, it felt more like a broader political shift. The time has arrived, and in two weeks, EU officials will meet to make a final decision on the Digital Omnibus and the amendments to the GDPR and the AI Act (among other topics). As I wrote in my newsletter, several of the proposed amendments weaken AI regulation in the EU and go against the protection of fundamental rights. If you are European, if you were hopeful for a Brussels Effect in AI, or if you are interested in the protection of fundamental rights, I invite you to read my full article below. - 👉 To learn more: - Read my article about the Digital Ominibus first draft and join my newsletter's 93,500+ subscribers (link below). - Join the 29th cohort of my Advanced AI Governance Training. Among the topics I cover in depth are the EU AI Act, the Digital Omnibus, and the European AI strategy (link below).

the economic collapse of gen z in four charts https://t.co/FoWNscQFPh

the economic collapse of gen z in four charts https://t.co/FoWNscQFPh

I love Twitter's new custom feed I just wish it wouldn't reset all the time to the default For You whenever I re-open the app. https://t.co/1U1zJubzK8

Day 7 of Building AGI for my Hermes Agent: I got my GPT-5.4 calls from 270x to 0x per day Today I found one of the dumbest parts of my setup where routine agent work was quietly burning GPT-5.4 token credits. Scanning, summaries, low-risk review, background thinking. These should not be hitting GPT-5.4. So I rebuilt the stack. Using @leopardracer’s Mac Mini local model setup, I got Local Qwen 35B A3B running for the always-on cognitive work as a test. Now the system works like this: - Local Qwen 35B A3B for scanning, summaries, low-risk review, and constant background thinking - MiniMax M2.7 for approved coding work - GPT-5.4 for final planning and high-judgment approval - No model for preflight checks when queues are empty The result: - GPT-5.4 cron calls per day went from 270 to 0. - Now GPT-5.4 only fires when there is real planning work or a deliberate escalation. Frontier intelligence is too expensive to waste on routine cognition, and my system got dramatically cheaper overnight, and much cleaner too. A lot of people are overspending because they are using the smartest model for the wrong jobs. Not every thought deserves a frontier token. Follow @gkisokay to see what happens next.

A video on how to install hermes agent because all of my friends and family keep asking me for a tutorial https://t.co/pYvInGx67N

This skill is actually cool af. Say Hi to Frank https://t.co/wcVUtPuWUC

This skill is actually cool af. Say Hi to Frank https://t.co/wcVUtPuWUC

I'm going all in on Hermes (@NousResearch, @Teknium1) as my entire agent and coding stack. Six profiles. One shared self-hosted memory store. Zero hosted-coder dependencies. The fleet: - pmax-mousa — my own WhatsApp + Email + Google Workspace agent - pmax-tarek — my co-founder's Telegram + Email agent - pmax-dareen — our content creator's WhatsApp assistant (LIVE on real client chats) - pmax-content — background content ops - pmax-ops-observer — daily health reports - pmax-coder — my primary coding CLI, no hosted coder, no gateway The model dial — this is the part I'm most excited about: pmax-coder runs on GLM-5.1 native via the https://t.co/s6oYqmfv05 Coding Plan (@Zai_org, quarterly $45). Direct to https://t.co/HUrIPiINWn, no middleman, no OpenRouter tax. GLM-5.1 published the exact thing I needed — a frontier coder at a flat price I can plan around. I've spent the last three days heads-down just getting the system running. Not tweaking it. Not optimizing it. Getting it to stand up end-to-end without a single load-bearing piece silently falling over. Six profiles, one memory store, two hosts, a dozen services, launchd, Tailscale, native provider pinning, patch re-application, ghost-process recovery, bridge port collisions, FTPS quirks, CI cycles, Qdrant lock contention, Happy Eyeballs hangs — every one of them a real bug I hit and fixed before I could move on. The three days are the story. The five gateway profiles (pmax-mousa, pmax-tarek, pmax-dareen, pmax-content, pmax-ops-observer) all run on qwen/qwen3.6-plus via OpenRouter native Alibaba routing (@Alibaba_Qwen, @OpenRouterAI). I pinned native-only with a strict provider.only patch so nothing silently falls through to a more expensive lane. Offline fallback everywhere is gemma-4-31b-it-4bit served by oMLX on the Mac Studio. If OpenRouter or https://t.co/s6oYqmfv05 goes sideways mid-conversation, every profile transparently fails over to local MLX inference and the user never notices. Swapping models is one YAML line. The real unlock: unified self-hosted memory. Every Hermes profile reads and writes one mem0 store on my MacBook (Qdrant + Ollama nomic-embed-text embeddings, zero cloud). Claude Code (@claudeai, @AnthropicAI) is wired to auto-broadcast every session turn into the same store via a Stop hook. The direction of flow is Claude writes, Hermes listens. Anything I decide in a Claude Code session is visible to the WhatsApp agents on my very next message. Nothing gets re-explained. Ever. Two-host architecture over Tailscale: MacBook (100.x.x.x) is the service layer. It runs mem0-server on 7437, task_server v1.1.3 on 7439, the guru-code router cache on 7450, the content-review webhook on 7438, all the Claude Code hooks, daily backup cron, and the mem CLI. Mac Studio M4 Max (100.x.x.x) is the agent layer. It runs Hermes v2026.4.13-118, all six profile gateways under launchd, the Hermes dashboard on 9119, the WhatsApp and Telegram bridges, and the Google Workspace OAuth session. Both hosts are pinned to IPv4 over the tailnet because macOS Happy Eyeballs was randomly hanging on IPv6 tailnet paths — one flag on every curl and ssh killed a whole class of flakiness. Huge credit to @brian_cheong — his push on idempotency-on-retries directly shaped task_server v1.1.3 (Idempotency-Key header on every write path), https://t.co/mrslfLodFL deterministic run_id dedup, and the guru-code router's response cache. Without that, retried agent actions would silently double-fire — a tool call would hit twice, a message would get sent twice, a file would get written twice. Whole classes of bugs I'll never write now. (btw I knew nothing abt idempotency bar — thanks dude) What else ships with the stack: Daily Qdrant and task_server backups with a 14-day rotation, plus a weekly full Hermes zip. Ghost-process immunity on launchd restarts (a startup_guard script kills any zombie https://t.co/eDFb9bsTOf holding the Qdrant lock before mem0-server boots). A native-provider pinning patch that wires provider_routing.allow_fallbacks straight through to OpenRouter. A secret redactor that runs on every Claude Code turn end so OpenRouter keys, Anthropic keys, GitHub PATs, and Bearer tokens can never leak into transcripts. A mem audit command that scans the memory store itself for leaked patterns. And a `fleet` one-shot status command I can run from any terminal to get a color-coded snapshot of every service on both hosts plus GitHub Actions status plus the Hermes patch inventory. Over these three days I also pulled 98 commits of upstream Hermes in two passes (70 + 28) without losing a single custom patch. An update-check cron inventories every local patch weekly so nothing regresses silently. Upgrades are safe. That's the invariant I wanted and I finally have it. None of this is a custom AI platform. It's Hermes doing what Hermes does, plus a few surgical patches I kept small enough to re-apply on every upstream pull. The whole philosophy is minimal lock-in: use the upstream as much as possible, patch only the load-bearing seams, never fork. The point isn't that Hermes beats every other coder tool today. The point is it's mine. I own the model dial, the memory store, the tools, the hooks, the backup policy, the security posture, the failover behavior. When something breaks I fix it. When I want to upgrade I upgrade. When I want to swap models I swap models. No middleman. No platform. No rug pull risk. Reports from the field to follow.

This is your sign to submit your #GitHubUniverse session idea ✨ Call for sessions is now open through May 1 👇 https://t.co/gRdMaPQ5NM https://t.co/KS7yYUuLx2

@thelumiereguy Should be fixed now thanks for the report! https://t.co/HqOI8uTBhq

速度真的好快！昨晚刚推荐，就被Hermes收入！ 感谢Hermes创始人！ https://t.co/zOl5IralEy

速度真的好快！昨晚刚推荐，就被Hermes收入！ 感谢Hermes创始人！ https://t.co/zOl5IralEy

@chipcamel @NousResearch Fixed and added, please `hermes update` to access :) https://t.co/WOurUGPuTG

Document OCR benchmarks are still an open problem Existing document OCR benchmarks are either too narrowly focused on a specific type (e.g. FinTabNet, ChartQA), or on documents that aren’t reflective of real-world tasks (e.g. OmniDocBench, OlmOCR-bench on over academic papers) ParseBench is a step towards solving this problem. * It tries to comprehensively cover real-world document distributions within the enterprise. * It contains comprehensive evaluations across 5 different dimensions (tables, charts, content faithfulness, formatting, grounding). * It tries to use metrics that optimize for agent semantic understanding rather than structural similarity. We released this yesterday, and there’s a TON of content: 1. Whitepaper 2. HF dataset 3. Github repo 4. Blog 5. Video And today, we’re excited to feature https://t.co/FYbk3s6M2w, our home page website for ParseBench 💫 come check it out! Take a look at some of our other materials if you’re interested: Blog: https://t.co/57OHkx0pQW Paper: https://t.co/Ho2oH2xEAM

How do you learn to trust AI? When it works even in a noisy environment. This is @typelessdotcom. Faster than typing. And you don’t need to turn down the music to use it. https://t.co/f2Oh0awxNb

Falcon 9 launches 29 @Starlink satellites from Florida https://t.co/LArWOnC3tD

+2 more