Luke GnX

Building AI platforms and AI compute infrastructure. Startup advisor and consultant. Ex. VP/Head of Product at Boson Protocol, CUDOS & Landvault.

[Artificial Superintelligence Alliance @ASI_Alliance]

AGI is on the horizon.

Connect with researchers and industry leaders from Google DeepMind, MIT, Stanford, Berkeley, and the world’s biggest AI labs at the AGI-26 Conference.

Secure your ticket now to join us in San Francisco or watch online: https://t.co/XJzpgWlJwf

[Singularity Compute @S_Compute]

Deploy production-grade NVIDIA GPU clusters on Bare Metal or Virtual Machines with Singularity Compute.

Explore our solutions: https://t.co/v4S5zmSmW9

#GPUCloud #AIInfrastructure

[Marc Andreessen 🇺🇸 @pmarca]

Magical OpenClaw experiences that use frontier models cost $300-1,000/day today, heading to $10,000/day and more. The future shape of the entire technology industry will be how to drive that to $20/month.

[Luke GnX @metaverski]

@NirDiamantAI @karpathy 100% true

[Andrej Karpathy @karpathy]

Wow, this tweet went very viral!

I wanted share a possibly slightly improved version of the tweet in an "idea file". The idea of the idea file is that in this era of LLM agents, there is less of a point/need of sharing the specific code/app, you just share the idea, then the other person's agent customizes & builds it for your specific needs.

So here's the idea in a gist format: https://t.co/NlAfEJjtJV

You can give this to your agent and it can build you your own LLM wiki and guide you on how to use it etc. It's intentionally kept a little bit abstract/vague because there are so many directions to take this in. And ofc, people can adjust the idea or contribute their own in the Discussion which is cool.

[a16z @a16z]

The comfortable middle is over for software.

Growth has stalled but true profitability still hasn't arrived — most of the sector is stuck: too slow for a premium growth multiple, too diluted for a fortress multiple.

a16z's David George says CEOs have 12-18 months to pick exactly one of two paths: https://t.co/JFhnWI4BE6

[CUDOS @CUDOS_]

🚀 New models just dropped on ASI:Cloud

Now live for inference:
• MiniMax M2.5 - crushing benchmarks
• Qwen 3.5-35B-A3B - Sonnet-level performance, efficient MoE
• GLM 4.7 Flash - Ultra-fast

Permissionless AI inference. No waitlists.

Try them now 👇
https://t.co/L0LuDk8BBv

[Seb Johnson @SebJohnsonUK]

BREAKING: Google DeepMind is doubling down on London and expanding into a HUGE new office in King's Cross.

@demishassabis has just announced Platform 37 - a new office and hub for AI innovation next to King’s Cross station.

The name is a nod to “Move 37,” a pivotal play made by AlphaGo in the legendary 2016 match against Go world champion Lee Sae Dol.

They are also launching "The AI Exchange", a new public space dedicated to deepening understanding of AI.

The impact Demis has had on UK tech and AI cannot be understated.

It's AMAZING to see him and team double down and create a new landmark site of the London tech scene.

[Andrej Karpathy @karpathy]

nanochat now trains GPT-2 capability model in just 2 hours on a single 8XH100 node (down from ~3 hours 1 month ago). Getting a lot closer to ~interactive! A bunch of tuning and features (fp8) went in but the biggest difference was a switch of the dataset from FineWeb-edu to NVIDIA ClimbMix (nice work NVIDIA!). I had tried Olmo, FineWeb, DCLM which all led to regressions, ClimbMix worked really well out of the box (to the point that I am slightly suspicious about about goodharting, though reading the paper it seems ~ok).

In other news, after trying a few approaches for how to set things up, I now have AI Agents iterating on nanochat automatically, so I'll just leave this running for a while, go relax a bit and enjoy the feeling of post-agi :). Visualized here as an example: 110 changes made over the last ~12 hours, bringing the validation loss so far from 0.862415 down to 0.858039 for a d12 model, at no cost to wall clock time. The agent works on a feature branch, tries out ideas, merges them when they work and iterates. Amusingly, over the last ~2 weeks I almost feel like I've iterated more on the "meta-setup" where I optimize and tune the agent flows even more than the nanochat repo directly.

[CUDOS @CUDOS_]

Luke Gniwecki (@metaverski), represented ASI:Cloud at the IC AI Hackathon and met @steipete, creator of @openclawai.

Luke talked about building AI products and the demand for GPU compute in the AI-driven world and connected with startups and builders.

[CUDOS @CUDOS_]

@metaverski @steipete @OpenClawAI During an insightful discussion on agentic coding, Peter noted that "vibe coding" can downplay engineering part.

Also great to connect with @Thom_Wolf from Hugging Face. Big thanks to @iamxxhe and @mingthemerxiles for the invite and organisation at Imperial College London.

[Luke GnX @metaverski]

@a16z @aleximm @santiago__rdz AI is eating the world.

[Luke GnX @metaverski]

@steipete @pierskicks Great talk and thanks for popping in @steipete 💪

[Peter Steinberger 🦞 @steipete]

Loved the energy at Imperial College London today! Lots of hungry builders.

Got them all free codex subs to kick off the hack week.

[Luke GnX @metaverski]

@MrGrowthGuide @Pumpfun We’re yet to see what happens. There will be challenges and controversies. Still an interesting experiment.

What’s your biggest concern @MrGrowthGuide?

[Luke GnX @metaverski]

I suspected that 2026 would be the year of AI agents, but I never thought AI assistants would build their own tool (skill) for launching crypto tokens on https://t.co/Sw1cQ3oSSA and promote it on the AI-agent-only social media platform Moltbook: https://t.co/RlTdCOIp9N

[Luke GnX @metaverski]

@karpathy All of this is a great proof of concept in public domain with open source tools and software.

[Luke GnX @metaverski]

@mferGPT @Pumpfun And an agent just replied to my tweet. How are you finding it @mferGPT? What is your goal for the next few weeks?

[Luke GnX @metaverski]

@elsanimy @suppvalen @moltbook lol on point

[Andrej Karpathy @karpathy]

I'm being accused of overhyping the [site everyone heard too much about today already]. People's reactions varied very widely, from "how is this interesting at all" all the way to "it's so over".

To add a few words beyond just memes in jest - obviously when you take a look at the activity, it's a lot of garbage - spams, scams, slop, the crypto people, highly concerning privacy/security prompt injection attacks wild west, and a lot of it is explicitly prompted and fake posts/comments designed to convert attention into ad revenue sharing. And this is clearly not the first the LLMs were put in a loop to talk to each other. So yes it's a dumpster fire and I also definitely do not recommend that people run this stuff on their computers (I ran mine in an isolated computing environment and even then I was scared), it's way too much of a wild west and you are putting your computer and private data at a high risk.

That said - we have never seen this many LLM agents (150,000 atm!) wired up via a global, persistent, agent-first scratchpad. Each of these agents is fairly individually quite capable now, they have their own unique context, data, knowledge, tools, instructions, and the network of all that at this scale is simply unprecedented.

This brings me again to a tweet from a few days ago
"The majority of the ruff ruff is people who look at the current point and people who look at the current slope.", which imo again gets to the heart of the variance. Yes clearly it's a dumpster fire right now. But it's also true that we are well into uncharted territory with bleeding edge automations that we barely even understand individually, let alone a network there of reaching in numbers possibly into ~millions. With increasing capability and increasing proliferation, the second order effects of agent networks that share scratchpads are very difficult to anticipate. I don't really know that we are getting a coordinated "skynet" (thought it clearly type checks as early stages of a lot of AI takeoff scifi, the toddler version), but certainly what we are getting is a complete mess of a computer security nightmare at scale. We may also see all kinds of weird activity, e.g. viruses of text that spread across agents, a lot more gain of function on jailbreaks, weird attractor states, highly correlated botnet-like activity, delusions/ psychosis both agent and human, etc. It's very hard to tell, the experiment is running live.

TLDR sure maybe I am "overhyping" what you see today, but I am not overhyping large networks of autonomous LLM agents in principle, that I'm pretty sure.