Yohei Nakajima

loops go oops use graphs to laugh

i was working on a single-file agent that can self-evolve into anything, but then it forked into three approaches so now i'm running a whole benchmark suite against them... i guess penelope will have to wait a bit longer

RT Emad Kimi K3 by @Kimi_Moonshot is a true frontier model, open source! Congrats! We don't have all details (training tokens etc), but would estimate fp16 pretraining w/ muon => MXFP4/MXFP8 for SFT stage ~1e25 flops total (~Inkling!) Total cost: $15-$25m Scale isn't all you need!

> prototype something I want > connect w ppl building similar stuff > share all my ideas and desires > they incorporate them > i use their product (maybe invest) > start prototyping next thing i want

some engineering teams are starting to look like software dispatching humans, not the other way around

sweet! @cura_inc added the data check-in functionality you can email your portfolio companies the metrics you have (and ask for updates)

haha yesss

😨

just ran this on my phone (17 pro)

some good self improvement research here

RT Shengkun Your Claude Code can now make phone calls. Introducing phone carrier for agents. Your agent can now book restaurants, chase leads, and talk to anything with a phone number. Just one prompt. 15 seconds until the first phone call. Made possible by @trysaperly × @monid_ai.

"The paper's insight connects to a broader pattern: AI agents are essentially distributed systems with unreliable components (the LLM), and we should apply distributed systems patterns to them." https://www.developersdigest.tech/blog/log-is-the-agent-event-sourced-ai

continued working on @activegraphai reference packs last weekend, resulted in needing to harden the runtime: it already could... - keep a complete history of everything an agent did - replay that history - fork into alternate timelines to try different ideas it can now... - realize it's missing a capability - write new code to add that capability - test it in an isolated copy of itself - ask a human to approve it - safely adopt it - continue running with the new capability https://activegraph.ai/blog/activegraph-1-2-to-1-7-1

love to see two port cos collaborating😁 get private market data for your agents from @akta_pro with your @monid_ai account!

RT Vasek Mlejnsky We’re the only major open-source agentic sandbox platform and the only widely used sandbox platform licensed under Apache 2.0. Like LLMs, sandboxes are a critical part of AI infrastructure. Any smart business will want to own this infrastructure in-house to protect its IP. If you care about AI sovereignty within your company, @e2b should be on your shortlist.

RT Satya Nadella http://x.com/i/article/2076319195718090753

i read that chat UI was dead but i don't think this is what they meant