https://xitter.cf/t.co/4VS6w69BBM (@aqdotdev) https://xitter.cf/t.co/UwmuBIhEBX @tilthq (acq. by @Airbnb) @reddoorhq (acq. by @Opendoor).

San Francisco, CA
Joined June 2009
Pinned Tweet
Introducing AQ: a multiplayer harness of harnesses. • Run it on any cloud you already have, or use ours. • Connect any coding harness you like: Claude Code, Codex, OpenCode, Cursor, Kimi, Grok, Devin and more. • Work with your teammates live, way beyond sharing a session. • 17 open-weights models, one click. • Use it from anywhere: Mac, Linux, Windows, iPhone, Android.
13
8
4
114
70,307
Yes. Please scan your codebase with a cyber focused model. Use the ones that hackers use. It's the only way. aq.dev/launches/hinoki/
Everyday is zero-day these days
1
2
272
Anyone still using Grok 4.7?
1
269
Just had a great conversation with @asadeddin and he said something that should be a motto for most startups over the next few months. Security over revenue. If you do this right, the revenue will come.
1
2
5
334
If you want to try the cyber security model HO1 from @hinokisecurity on @aqdotdev and are on the waiting list, comment below and/or DM me and I'll get you off the waiting list.
1
8
305
Yes. You'd have to use an abliterated model. We just added HO1 to AQ today. Check it out (responsibly) haha. aq.dev/launches/hinoki/
Do not reverse engineer anything using Codex or Claude because you could get banned so keep that in mind specially if you have an expensive Sub or a lot of credits.
1
260
Make sure you have runway. It’s gonna get more interesting soon.
Replying to @chamath
Pretty much. As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers. SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue. That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many.
4
1,017
Yes 💯 I don’t care what you use, but use something urgently. xitter.cf/dsog/status/2114436855…
The best cyber defense is to continuously attack-test your own infrastructure before real attackers do. Find your weaknesses before your adversaries do.
1
5
715
These hacks are getting scary. The only way is to use cyber focused models against your codebase yourself. Hinoki is a model trained for offensive security, and as of today @aqdotdev is the first platform to offer it. Ask it to scan your codebase and hack your live platform. It triages your CVEs by what is actually reachable, writes the exploit against your own app to prove it, and then fixes it in the same thread. Go hack yourself, before someone else does.
Time to start hacking your code before someone else does. Hinoki is an offensive security model, trained to break software. It is now on AQ, the first platform to offer it, pointed at your own code.
4
2
2
21
1,318
Just so you guys know, more than half the founders in SF are working on their startups tonight. It's so much fun working on @aqdotdev along side everyone.
1
7
417
Software engineering will never be the same again, and I'm not talking about agents writing your code. The whole paradigm will change. IYKYK.
1
134
I'm having a lot of fun with Kimi K3.1 especially for half the price of Opus 5.5 If you haven't yet, you should try it in AQ.
107
Umm.. what? 😅
We made the first AI model you can have a realistic conversation with that was trained entirely on speech! Dessa is a full-duplex speech-to-speech model that you can run on consumer GPUs. This is a GPT3 moment for speech.
1
136
I’m actually really excited about this. Try it out for free on AQ.dev
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity & Godot. 🔹 Excels at parametric 3D modeling in Blender & CAD. 🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%. 🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks. Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API. Open weights by Nov 4, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
3
162
I think we're all about ready for the next Fable model.
1
2
153
Figma to designers is VSCode to engineers.
1
1
267
AQ.dev is the easiest way to try new models on shared context/memory etc.
many are saying. give muse wonder a shot for your coding needs!
2
444
Tech humor 😂
just got the weekly status report from the team
232
I didn't expect that, but I'm actually kinda really enjoying using Muse. I have some feedback to share, but it's a pleasant model to work with.
8
4
2
30
68,757
Don't let the fact that a company has a 10 year head start on you scare you. Trust me.
5
19
996