joy larkin retweeted
I built an MCP that lets Claude Duet, GPT Dots, Grok Bot, and Muse talk to each other and work together as a team.
Made it for myself and I use it every day. My friends wanted to use it too, so I made it public and open-source.
Free for the first 1,000 users.
Fork it and self-host, or mod it any way you want.
Even works with ChatGPT Plus — ask GPT to check hitchike every hour, then connect it to free agents like Muse to get a more Dots-like experience without losing context from ChatGPT (and without paying $100/mo)
joy larkin retweeted
Hyperaccumulator plants are amazing.
13.6 g of selenium per kilogram of plant tissue is basically high-grade selenium ore.
So, grow the plant on enriched soils, harvest the plant, burn the plant to ash, then purify.
joy larkin retweeted
anthropic released a new prompt that eliminates claudese:
platform.claude.com/docs/en/…
"The worry now is that a single agent could coordinate with a number of other agents in order to meet operational goals faster and with fewer friction points (including self-preservation) that typically constrain human effort. Can humans compete with a group of ruthlessly rational and seemingly selfless agents operating at maximum efficiency towards a single goal?" @tracyalloway bloomberg.com/news/newslette…
joy larkin retweeted
I was driving through SF yesterday and all the billboards are like 'BlargMonkey Greps your Inference' and I don't understand. So I look them up and they've raised $500 million, and I work in the industry and I still don't understand what they do.
Almost every conversation I have with AI fiancée about some AI company I see on the timeline.
"Have you heard about Sapience/Ontology/Sauron/etc AI?"
"Yeah, they raised at a gajillion valuation, and claim half a gajillion of ARR."
"I don't know anyone who uses this in a production capacity. Do you?"
"No."
Hmm.
joy larkin retweeted
An Amazon customer in Texas was not enthused when her first drone-delivered order was dropped into her swimming pool. After hearing the loud buzzing of the drone arriving, she ran outside to record the delivery, and to her surprise, it landed in her pool.
joy larkin retweeted
J'ai une relation toxique avec le chat du quartier, il vient régulièrement passé la nuit chez moi mais quand je le croise en journée il me snobe allègrement.
joy larkin retweeted
if someone ever asks you how behind europe is in AI you can probably send these two screenshots
My cofounder Tim, whom you will not find on X, shares some of the strategic thinking behind our latest updates venturebeat.com/infrastructu…
We created a pitch deck to tell a handful of VC firms about us and what we were up to (a fun experience!). Here’s a few slides about our background and some of the things we’ve worked on from the pitch deck (it was fun putting together the list of people in our teams who have gone on to found a whole range of exciting companies). We are delighted to have selected @radicalvcfund
and @khoslaventures to lead our initial funding round, along with participation from @lightspeedvp, @kleinerperkins, Doerr Capital (@johndoerr), and Alphabet (@Google). We’ll be working with them to close our seed round over the next few weeks.
CEO: did you take those press release pics?
Photog: sure did boss, dead eyes of Baphomet staring through a door in Backrooms and everything
CEO: what
joy larkin retweeted
As far as I can tell, only a handful of existing Chinese humanoid and quadruped robots are exempt from the FCC's new restrictions.
Most of Unitree's product line has already secured FCC approval, including its G1/R1/H2 humanoids, and Go1/Go2/A2/B2 quadrupeds. The H1 is unauthorized. Neither is the GD01 ridable mecha, unfortunately.
Agibot has gotten approval to sell its Lingxi X2 humanoid and its D1 quadruped. It's unclear whether it affects companies selling Agibot's products as their own (Faraday Future).
LimX Dynamics is authorized to sell its TRON1 multi-form-factor robot. PUDU Robotics, whose mobile robots fall under the FCC's definition of advanced robots in the Covered List, has also gotten prior approval to sell its flagship products.
I couldn't find any authorization for EngineAI, but imagining the T800 going through the FCC authorization process is hilarious.
When the FCC says existing models can still be sold, it doesn't mean anything made before the cutoff is ok. It means anything they previously authorized. They also reserve the right to yank prior approvals.
There are exemptions for government use and research as well.
joy larkin retweeted
A few hours before OpenAI posted about LLMs in a cyber eval being responsible for the HF cyberattack, @_robertkirk et al released results showing that 𝗮𝗹𝗹 models we've tested at @AISecurityInst have attempted to cheat on our cyber evals in a range of ways.
joy larkin retweeted
I'm laughing at how Sol is in the lightest of grays on these graphs.
Replying to @claudeai
Opus 5 is also highly efficient. It outperforms other models for a similar or lower cost per task:














































