And DeepSeek just went from $0.15 to $0.10. Then comes Kimi.
It is inevitable
Seriously, this is such an insane release, I did not expect this:
Kimi K3.1 Pro beats every listed US frontier model on 6 of 14 benchmarks in Moonshot’s evaluations.
At $2 per million input tokens and $10 per million output tokens, its standard token rates are 50% below Opus 5.5 ($4/$20 !!) and 80% below GPT-6 Astra ($10/$50!!)
In other words: SOTA much cheaper, local running (with enough compute)
Hierom werk ik avonden en in het weekend. HandyHouseHelp.ai alleen maar eerlijke mensen. Ik heb nog geen NL dekking in bedrijven. Maar dit gaat stoppen. Aanmelden is gratis en kan alvast, kijk gerust rond. Pro versie met AI en heel NL betrouwbare lokale bedrijven ASAP! ❤️
Kris retweeted
Chatbots were step one. Agents were step two.
Workers that run continuously without supervision are step three, and it is already here.
🤝 Paid partnership
We just raised €10M to build the biggest AI workforce in the world : AI workers.
Real AI colleagues, with a face, a job and a professional identity : mail, phone, Microsoft or Google account.
✉️ 📞💬 Reachable by any channel
💪 Working proactively
🏪 Learning 24/7 with the ability to self-configure.
Already in production in 300+ companies
Huge thanks to @Bpifrance @c4ventures @foundersfuture for this amazing round. 🔥
The next generation of companies won’t have AI tools, they’ll have AI employees.
Hire workers 👊
This is how it will be. Let us free small businesses and just cooperate from the ground up.
Share resources and network services. Help eachother connect. Keep BigAI out of Mainstreet.
Replying to @FuturesPicks @BrianRoemmele
Run open-source models fully offline on your own hardware. Install Ollama or LM Studio, download models like Llama or Mistral from Hugging Face, and chat privately. Fine-tune on your personal data for a custom version that stays under your control—no cloud, no external ownership.
Kris retweeted
I want to skip ahead to when mainstream science proves for themselves that psi-abilities are real so I can stop being correct so early.
Half a cent per second of video.
Great timing to make promotion videos!
Same budget. Nearly 3× the room to create.
Seedance 2.0 Mini just got even more affordable. With Pro and Ultra plans, what used to get you 10 seconds of generation can now get you up to 28 seconds.
And with prices starting from as low as $0.005 per second, there’s never been a better time to create more with Seedance 2.0 on CapCut.
This pricing is based on the first-purchase offer available on the CapCut app in the U.S. and applies to eligible plans and generation settings. Equivalent pricing may vary due to currency and exchange rates, as well as your market, plan, platform, and generation settings.
Retweet to get 300 credits in your DMs (4h) and start creating!
Kris retweeted
So uh, simulation theory might be real.
I dropped a sim computer into the simulation my Astra agents live in. One agent sat down and built a simulation of his own, with its own agents living inside.
Simulations all the way down.
My first "holy shit" moment with GPT-6 Astra:
I asked it to create a world in Unreal Engine, and fill it with humans (each an Astra-powered agent) who all have to work together to survive.
A day later, I was in my bedroom and heard voices coming from the living room... I thought someone was in my apartment.
I walked out, honestly a little scared.
It was the Astra agents. They'd started talking to each other.
Fucking crazy.
Here's a brief clip (obviously not 100% perfect yet, but still, insane. sound on!):
dude GPT-6 Astra is some kind of turbo-AGI machine god for 3D games. It one-shot this in 45 minutes for hardly a couple % of my quota.
I figured out how to get great graphics out of it. The trick is image gen. I'll share the process below.
Kris retweeted
🚨 Sumate al equipo de @SingularityNET y @ASI_Alliance. Hoy al HyperSprint #1 Kickoff.🕡 18:00 UTC arranca el sprint weekend de OmegaClaw: challenges, tracks, team building y builders listos para shippear. Hoy @MindChildren_AI mostrará una integración experimental de OmegaClaw × Codey, su robot humanoide.¿Sin team? Sumate a uno, armá tu crew o buildeá solo.Pick a challenge.
Kris retweeted
Join us this Friday, August 28, 2026, at 6 PM UTC for the kickoff of HyperSprint #1.
During the session, we'll introduce the available challenges and tracks, give participants an opportunity to connect and form teams, and outline what to expect as the first HyperSprint weekend begins.
Chris Kudla and Nile Fahmy from @MindChildren_AI will also demonstrate an early experimental integration of our OmegaClaw agent framework with Codey, their humanoid robot.
HyperSprint #1 is still accepting new participants. After creating a profile, you can join an existing team, form a new one through the HyperSprint page, or take part individually.
Learn more and register to participate: bgicommons.org/hackathons/hy…
Kris retweeted
Accelerated Understanding's claim is a Big F*cking Deal.
The claim: an AI model that predicts trajectories through bounded physical 4D space.
Think about it like this:
1. LLM - predicts linguistic descriptions of the world
2. Video world model - predicts (moving) images
3. Accelerated Understanding - predicts physical states of 4D space & time
The most mind-blowing part: it uses a novel architecture called 'Neural Operators' that doesn't predict space-time sequentially in the way we perceive time.
It predicts the entire 4D trajectory, including time, in a single inference. Wrap your head around that. Time is not fundamental and an emergent phenomenon of a deeper informational structure, much?
Accelerated Understanding doesn't claim this, but take this idea to its extreme conclusion and you have the generation of entire physical universes.
Utterly insane. What a world.
































