There are now sufficiently many different examples of Erdos problems that have been resolved with various amounts of AI assistance and formal verification (see github.com/teorth/erdosproblem… for a summary) that one can start to discern general trends.

Broadly speaking, we now see an empirical tradeoff between the level of AI involvement in the solution, and the difficulty or novelty of that solution. In particular, the recent solutions have spanned a spectrum roughly describable as follows:

1. Completely autonomous AI solutions to Erdos problems that are short and largely follow a standard technique. (In many, but not all, of these cases, some existing literature was found that proved a very similar result by a similar method.)

2. AI-powered modifications of existing solutions (which could be either human-generated or AI-generated) that managed to improve or modify these solutions in various ways, for instance by upgrading a partial solution to a full solution, or optimizing the parameters of the proof.

3. Complex interactions between humans and AI tools in which the AI tools provided crucial calculations, or proofs of key steps, allowing the collaboration to achieve moderately complicated and novel solutions to open problems.

4. Difficult research-level papers solving one or more Erdos problems by mostly traditional human means, but for which AI tools were useful for secondary tasks such as generation of code, numerics, references, or pictures.

(1/2)

AI contributions to Erdős problems

A community database for the problems on the erdosproblems.com site - teorth/erdosproblems

^GitHub

What is NVIDIA without AI?

NVID.

Random neurodivergence thought: Stereotypes are often very binary.

People with ADHD are either loud boys who can't sit still for a moment or normal people who are a little unorganized.

Autistic people are either nonverbal and obsessing over trains or normal people who are a little awkward.

People with schizophrenia are always raving and hearing voices though. They're never normal people who are a little spacey. That's how stigmatized it is.

I'm having the first Radicle CI office hour this Friday at 08:00 UTC at meet.ffmuc.net/radiclecioffice…

If Radicle CI, distributed CI, or just Radicle, or just CI, interests you, feel free to join.

#Radicle #RadicleCI #OfficeHour

Jitsi Meet

Join a WebRTC video conference powered by the Jitsi Videobridge

^{meet.ffmuc.net}

2026 will see #deMicrosoft trending - what's your pick?

#NoAI #MS365 #DigitalSovereignty #OpenSource

Not ready to pay increased prices for AI? Choose alternatives: tuta.com/blog/microsoft-365-pr…

Final Cut Studio used to cost $1,300 at one point, and now it's available to students for $2.99 a month. That's a huge win in my books.

I was once, many moons ago, a student paying top dollar for Final Cut Pro, later managing a Final Cut Studio …uh… studio.

My #ScottAdams Story

In the 1990s, I worked as an office temp. I logged a lot of hours in a lot of different offices, and I had an instant and accurate way to sense how dysfunctional and toxic a workplace was as soon as I walked in.

I took note of how many #Dilbert comics were pinned up, and where.

If I saw one or two #Dilbert comics scattered around, I knew people had their gripes and complaints about their co-workers, but it was nothing too serious.

1/9