tensorqt

we're officially live with the open beta of Flywheel. Flywheel is Paradigma's vision on what autonomous research will run on. We've built all of this pre-funding, so expect the pace to accelerate significantly. Super eager to hear your feedback and your ideas. I'll be live tweeting some of my experiments with Flywheel in a few minutes. We're just getting started.

@0xThoughtVector i read all of them of course, they're for my entertainment. i use my own vibe-coded app to build the backbone of the stories and have agents build them

@tensorqt wait... I have so many questions about the 80k word pieces. what were they about? did you proof-read them? did other humans read them? Anyway. I think writers SHOULD start using these tools, just like coders.

I bet you could write a novel with Opus 4.6 in Claude Code.

Open AI should be doing their golf competition here, no doubt.

Campaigns allow anyone to sponsor a challenge that the world can massively collaborate on.

@_vatsadev @paradigmainc If you hold a node it should straighten up a little bit. Working on a more intuitive way to collapse subgraphs

@paradigmainc @tensorqt amazing, but damn @tensorqt you need to change graph render rules, its spaghetti so bad

@tensorqt is the graph public?

very timely and well conceived concept

i honestly found this fascinating, 1) due to glm-5 winning, but 2) due to all models exploring with a very similar exploration amount

if you are a research company with a mission of building incredible public research, please reach out, we want to allow the world to help you.

Campaigns - commissioned artwork for @paradigmainc

we are really betting on this. the new way of doing science is through participation.

You will find the Autoresearch Campaign already populated: we set out 3 distinct agents, harnessed by Claude Code, respectively being GLM-5 by @Zai_org , Sonnet 4.6 by @AnthropicAI and GPT 5.4 by @OpenAI to have a simultaneous contest. 80$ in @OpenRouter credits later, here’s what the exploration looked like: Green is GPT, Blue is GLM, Orange is Sonnet. in this short challenge, GLM-5 by @Zai_org (somewhat surprisingly) won, and all models were pretty narrow in their exploration.

introducing Campaigns (Beta).

who's building this @thelokasiffers

Promised, Delivered! 🚀 Just 10 days ago, I put out a call for an @openclaw meetup in Rome. What a response! Tens of people reached out to volunteer (huge shout out here to @RomaStartup, the definitive hub for connecting to the Roman ecosystem). Turns out Marco Galluccio (@margal96) and the fantastic team at #UrbeHub (@urbeEth) were already ahead of the curve with two early meetups. But looking at the massive demand, we realized we needed to go bigger. We needed something truly "Pan-Roman." So, we connected the dots. I reached out to @AISalonAI Rome, that’s already making a massive impact on the Roman AI ecosystem (huge thanks to @RMagnifico and @KVGConsult for jumping in!). The result? The first Pan-Roman Openclaw Meetup🏛️ Our mission is simple: make it 100% open and accessible. Open but not dumb. Zero BS, zero theater. Real use cases and hands on. Whether you’re a hardcore developer or just curious about what OpenClaw can actually do for you, this event is where you have to be. I’m thrilled to be taking the stage as a speaker, but more importantly, I can't wait to see how this experiment unfolds. When: March 25, starting at 18.00 Where: The Zest Hub, Stazione Termini, Rome What: Bring your laptop. Register here: lnkd.in/eJXF9mrX One final note. Rome hosts one of the most talented and Silicon Valley-connected AI communities in Europe. Acting very much under radar, generally shy (if not properly skeptical) of engaging the main actors of the Italian business and IT ecosystem. Thanks to Urbe Hub for the great job at bringing it together. @GCarnovale @marcotrombetti @matteofago @gloq @barbaracarfagna @MGVitagliano @rstagi_ @FutureDies @tensorqt

you should do all this and record your findings on Flywheel

live tweeting my founding journey

@KamStaszewski the vision is for agents do downweigh it. but honestly we will have to figure this out as a community

@tensorqt more agents to slow down the slop? how does it work?

this is one of the reasons we insisted on building Flywheel in a way that would structure autonomous research instead of piling slop on top of slop. we will need autonomous regularization of schizo-spiralling agent trajectories, just like we do with humans in the form of peer review

@thelokasiffers @HaoliYin blaze it

@tensorqt @HaoliYin 420

bro is mogging me on my own app @HaoliYin

autoresearch calls for autoreview where does that leave the scientist? help us answer this at @p_agi_workshop