Bloc97

Hermesmaxxing Thinking of maybe adding live notifications so you can see the responses stream in on the lock screen.

I said I'd update if they follow through, and now, Anthropic has granted us security privileges with Claude.

Hey all security people, am looking for comments and suggestions on this PR, if you have constructive thoughts or comments please check the PR out and let me know as we build towards far improved security for secrets: github.com/NousResearch/h…

Our database and data engineering expert @yoniebans made some major improvements to the way sessions are stored and accessed. This will save something like 20-40% of the disk space used by Hermes Agent to operate, speed up session loading, and overall makes the codebase cleaner, simpler, and better architected! `hermes update` to access early or wait for the next major release :)

Subword boundaries are the second meaningful effect. Adding end-of-subword markers as input embeddings produces a large gain throughout training (H3): end-boundaries leak future bytes (whitespace always follows an end-boundary, for example) and simplify the next-byte prediction task. Start-of-subword boundaries cannot leak the future, and they also help. When start-boundaries are provided only during the first 50k training steps and removed thereafter for both training and validation, the improvement persists; end-boundaries do not survive the same intervention. One reading is that start-boundaries supply a morphological inductive bias (H4), while end-boundaries supply a near-term prior the model becomes dependent on.

Today we release a study on decoupling the benefits of subword tokenization for language model training, by simulating each suspected benefit one at a time inside a 1.7B byte-level pretraining pipeline. We formulate seven hypotheses for why subword LLMs outperform byte-level LLMs (covering computational efficiency, structural priors over subword boundaries and positions, and the optimization objective) and implement each as a controlled intervention against a byte-level baseline. Three of the seven move the validation loss at this scale; the rest either have negligible effect or hurt. Validated at 1.7B parameters on fineweb-edu with a LLaMA-3 architecture, with 68M-parameter replications in the appendix. The work was led by Théo Gigant, Bowen Peng, and Jeffrey Quesnelle. Paper: arxiv.org/abs/2604.27263

$VVV is now available to trade on Robinhood.

Meet Gemini 3.5 Flash — our strongest agentic and coding model yet. It delivers frontier-level performance at 4x the speed of comparable frontier models — often at less than half the cost. Generally available, starting today. 🧵 #GoogleIO

People are generating over 1.5 billion images a week in ChatGPT. Researcher @kenjihata joins Product lead @adele__li and host @AndrewMayne to explore the new use cases and trends emerging since the launch of Images 2.0.

Introducing Composer 2.5, our most powerful model yet. It's more intelligent, better at sustained work on long-running tasks, and more reliable at following complex instructions. For the next week, we’re doubling the included usage of the model.

Fast mode now defaults to Opus 4.7 in Claude Code. Try it out today with /fast

1/ The new $sGHO is live. Here's what changed across the product, and what it means if you're currently holding legacy stkGHO ↓

The new and improved Savings @GHO is live. Instant deposit and withdrawals with a 4.25% Aave Savings rate. One of the better stablecoin rates in the industry.

.@AerodromeFi V2 & V3 pools are now accessible directly on KyberEarn. One place to discover pools, analyze performance, and decide where your liquidity goes next. More pools. More ways to earn. 👉 kyberswap.com/earn/pools?cha…

You've been asking for this one... Now in preview: Codex in the ChatGPT mobile app. Start new work, review outputs, steer execution, and approve next steps, all from the ChatGPT mobile app. Codex will keep running on your laptop, Mac mini, or devbox.

parameter golf was a blast. 2,000+ submissions. 1,000+ verified github accounts. ideas ranging from quantization and depth recurrence to TTT LoRA, SSMs, H-nets, JEPA, and more. autoresearch made iteration dramatically faster — and led to emergent bulletin boards, issue threads, unofficial leaderboards, and agent-built writeups that helped everyone learn from everyone else. it felt like a glimpse of where interaction with AI is headed: humans setting taste and direction, agents helping explore, coordinate, and share what works. our goal was simple: make ml research accessible to anyone, anywhere. it was amazing to see that happen. full recap: openai.com/index/what-par… future events: jobs.ashbyhq.com/openai/form/op…

Another reason to switch to Codex.

@FlamePanda87 @nyaraVT You're pushing 40 years old. Time to grow up.

@ImWatson91 @nyaraVT

The thing is.. I'm not a man. I'm okay with having the same risks that estrogen has in cis women. I'm okay with breast growth. I'm okay with infertility. I hope this helps.