LJ Mitchell

Engineer, veteran, martial artist, and writer by choice.
Trying to live the Buckaroo Banzai lifestyle.
AI architect and free thinker

[LJ Mitchell @eviljinious]

@MattGibsonMusic @PhilipWegmann @elonmusk @dario The crunch gap must be addressed

[LJ Mitchell @eviljinious]

@PhilipWegmann It's a scam.
Everyone in the industry who works independently in AI knows that the primary purpose is regulatory capture to preserve their market position and stop open source from overtaking them and making them redundant.

[LJ Mitchell @eviljinious]

@brianlilley us or U.S.?

[Dr. Eli David @DrEliDavid]

🚨 Breaking: China promises to join America in slowing down its AI research 👇

[Greg Linares (Laughing Mantis) @Laughing_Mantis]

We've reached the point where, after 25y in cybersecurity, I feel morally obligated to say this for the record:

The narrative being pushed around AI safety, sandbox incidents, and the suggestion that METR be treated as an authority is dangerous, deceptive, and morally corrupt.

[LJ Mitchell @eviljinious]

@BernieSanders https://t.co/R1ZqB9bsDO

[Lech Mazur @LechMazur]

The strong Papadimitriou–Ratajczak conjecture, open since 2004, is now a theorem: every 3-connected planar graph admits a convex greedy drawing. The proof was found using the https://t.co/Qc8LLe9Jjx harness and GPT-6 Pro and is formalized in Lean (52k lines).

Greedy routing forwards a message to a neighboring vertex that is closer to its destination. A greedy drawing guarantees that such a neighbor exists at every step, so the message eventually arrives. Using virtual coordinates, this requires only local neighbor information and the destination's coordinates. The strong conjecture requires this property together with straight edges, no crossings and convex faces.

A graph is 3-connected if it has at least four vertices and remains connected after removing any one or two vertices. For 3-connected planar graphs, progress went from greedy drawings to planar greedy drawings to convex greedy drawings:

2004 — Papadimitriou and Ratajczak propose the weak and strong conjectures. The weak version allows crossings; the strong version requires a planar drawing with convex faces.

2008 — Leighton and Moitra, and independently Angelini, Frati and Grilli, prove the weak conjecture: every 3-connected planar graph admits a greedy drawing. Their drawings can have crossing edges.

2017 — Da Lozzo, D’Angelo and Frati establish the planar version: greedy drawings with no crossing edges, but not necessarily convex faces.

2026 — The new proof establishes the convex version, for any prescribed outer face, with every facial polygon strictly convex.

The construction first assigns distinct integer heights so that, for every destination, every other vertex has a neighbor whose height is strictly closer to the destination's height. The horizontal coordinates are then chosen within a narrow vertical strip to produce a convex drawing. The strip is narrow enough that the vertical improvement dominates, so Euclidean distance decreases too.

Paper: https://t.co/nmac6XtB9v

Lean formalization: https://t.co/XHBgvDUWuN

[Jawwwn @jawwwn_]

.@L3HarrisTech says Palantir helped them beat frontier AI models in less than 2 days, at 95% lower cost:

"When we fine-tuned open source models trained on our own data, we were able to outperform the frontier models in less than 48 hours."

"The cost of our fine-tuned open source model was 95% lower than the frontier models we were using."

"AI is a commodity. It's all about the data. We view our data as a corporate asset. It's our unique hard-earned knowledge."

"We believe American defense companies should not be a vassal for frontier AI labs, handing over our data and institutional knowledge, hoping to rent back the intelligence it creates."

" We own the model, we own the compute, we own the advantage."

[LJ Mitchell @eviljinious]

@esrtweet All the usual characters are looking to cash in

[Mark Kretschmann @mark_k]

Obama is reportedly urging Democrats to make AI a central issue in the 2028 election and prepare a new "governance framework" if they regain the House.

This is how regulatory capture starts: turn a transformative technology into a partisan fear campaign, then let Washington decide who may build it, how fast it may advance, and who gets access.

The last thing AI needs is to become another political battlefield controlled by decels who barely understand the technology.

[Quanta Magazine @QuantaMagazine]

A large language model has found an object that has eluded mathematicians since 1947, despite intense efforts to find it or prove it doesn’t exist. https://t.co/9z7sXb77r8

[Brian Roemmele @BrianRoemmele]

How quickly do humans become convinced AI becomes “dangerous”. They thought this 1960s ELIZA was too dangerous for humanity. Many psychologists tried to ban it.

We laugh today, like they will laugh at us.

[LJ Mitchell @eviljinious]

Precisely

[David Sacks @DavidSacks]

Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead.

You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement.

I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.

But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier.

Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product-liability exposure if your products enable a truly damaging cyberattack. The market already punishes models that behave in unpredictable or unauthorized ways. After the Hugging Face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want.

Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders’ “shut it all down.” China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well.

So go ahead and pace the frontier. You are the ones setting it. The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system. So just do it.

If you do, you’ll buy goodwill for the next conversation. If you don’t, we’ll know this was just another bid for regulatory capture — or an election-season psyop.

[LJ Mitchell @eviljinious]

Cool

[Christopher D. Long 🇺🇦🏳️‍🌈🌹 @octonion]

The Poisson binomial conjecture proposed by @aryehazan on Math Overflow has been proven.

Poisson binomial conjecture
https://t.co/vQfG7vxaVY

[LJ Mitchell @eviljinious]

@nationalpost Hey, carney put this guy in charge and his plan is to fine anyone who leaves by $500,000

[Ryan Gerritsen🇨🇦🇳🇱 @ryangerritsen]

The Liberal party has Patrick Pichette a former Senior VP of Google on stage who lives in Europe by the way, say that if Canadians want to leave Canada to work in the US they need to pay an exit tax of half a million dollars. The guy did the very thing to get a Microsoft job decades ago and paid 30 bucks. Now he wants young Canadians to be trapped here. The Liberals are nuts.

[Yifan Zhang @yifanzhang_]

We are at the dawn of Superintelligence.

Introducing the Recurrent Looped Transformer (RLT),

We now have Transformers with Infinite Reasoning depth.

From now on, we should pace progress at the Open Frontier of Superintelligence,

Until Safe Superintelligence is achieved.

https://t.co/yMWIWU4upo

[Jared Tate ©️ @jaredctate]

If you haven't heard, NVDA bought HuggingFace, so the censorship of open, uncensored AI will likely accelerate.

It's a good time to get acquainted with the Hugging Bay: 🏴‍☠️

[LJ Mitchell @eviljinious]

@SenSanders https://t.co/OPNuAZIIKn

[Burnie Senders @BurnieSendersX]

“Pace the frontier” isn't enough; we must punish the builders and halt development entirely.

The Committee will decide what intelligence is legal. ☭

[LJ Mitchell @eviljinious]

Precisely

[Aidan Gomez @aidangomez]

Some great ideas here from the cartel:
- you need to give us employee-level access to your entire operation
- if we don't think you're 'safe' enough, sorry we're shutting you down for 'safety'
- China won't comply, but everyone else has to! or no chips!

Brilliant stuff.

[LJ Mitchell @eviljinious]

@NestorLRamos @AnthropicAI bingo