Rich Stureborg

CEO @Acrebase. Board member @DukeU. NLP PhD

[Rich Stureborg @rstureborg]

The terms for this sucks, and you agree to them by just applying.

- §2.4 Anthropic "may develop similar or competitive technology"
- §2.4 permits reuse, for any purpose, of general ideas, know-how and techniques remembered by personnel looking at your data
- §D.4 prohibits accessing their services to build competing products or services
- §F.1 allows them to retain and perform safety reviews, superseding ZDR (whats the point of ZDR then?)

and it could get worse later!
- §6 permits changes to the addendum at any time without notice or liability

@jason you advising your start ups to apply for this?

[1] https://t.co/IFeoVixOsb
[2] https://t.co/PKJFl6UBlr

[Rich Stureborg @rstureborg]

@natolambert 🤔 This was how people felt about coding agents for a long time. That "where's the shovelware" article went around a lot back then, with skepticism of LLM impact on coding.

I think people overlook the impact of diffusion. We see capability instantly impact takes a few months.

[Rich Stureborg @rstureborg]

@natolambert This article, for the uninitiated:
https://t.co/gBDaS5wcqq

[Rich Stureborg @rstureborg]

@natolambert 🤔 This was how people felt about coding agents for a long time. That "where's the shovelware" article went around a lot back then, with skepticism of LLM impact on coding.

I think people overlook the impact of diffusion. We see capability instantly impact takes a few months.

[Rich Stureborg @rstureborg]

So insanely high value, can't believe she's sharing this for free.

[Elizabeth Yin 💛 @dunkhippo33]

1/ Sharing our deal stats for September 2026:

1,922 companies submitted pitches to Hustle Fund last month. Here's what we saw, including why we passed, what valuations looked like, and what founders are building. Read on >>

[Rich Stureborg @rstureborg]

@SKatalystAI Ya I think the key is just to realize anthropic isn't going to act like a VC / typical accelerator. They're another company and may be a competitor. Always worth knowing the terms

[Rich Stureborg @rstureborg]

@SKatalystAI @spencerclarkdev agree!

[Rich Stureborg @rstureborg]

So simple and elegant. This is S tier UX design. AI is not (yet) capable of this type of thing.

[Alejandro Buján @alejandr0bujan]

I made my Mac hang every screenshot on a clothesline

[Rich Stureborg @rstureborg]

@omarsar0 the "-as-a-Judge" is redundant

[Rich Stureborg @rstureborg]

The repo summarizes some of this but here are 2 of the more important problems they made progress on:

1. Quasi-Riemann hypothesis: This gives much stronger guarantees on how irregularly primes can be distributed, by restricting where zeros of the zeta function and related functions can occur. That said, this establishes a boundary at 7/8. the full Riemann hypothesis requires 1/2 and remains unsolved.

2. Unique Games Conjecture: New limits on how well polynomial-time algorithms can approximate optimization problems like Max-Cut and Vertex Cover. The caveat on this one is that interpreting it still assumes P /neq NP. Doesn't resolve P = NP.

[zack @zack_overflow]

Of course, that’s your contention. You’re a first-year machine-learning engineer. You just got finished readin' Attention Is All You Need, probably watched a Karpathy video too. So now you’re convinced everything is just transformers and scaling laws.

That's gonna last until next month when you discover convolution, and then you're gonna be talkin' about inductive biases and locality and how CNNs were actually incredibly compute-efficient for vision.

Then you're gonna read the FlashAttention paper, and suddenly everything's about IO complexity and SRAM and how FLOPs don't matter because the whole goddamn thing is memory-bandwidth bound.

That'll last until somebody shows you an MoE model, and then you're gonna be in here regurgitating DeepSeek, talkin' about expert parallelism and active parameters and how dense models are economically obsolete.

Then six months from now you'll write one shitty Triton kernel, look at an Nsight trace for the first time, and start telling everybody Python isn't actually the bottleneck because the GPU is asynchronous.

And by next year you're gonna be standing right here explaining to me that your 400-billion-parameter model is fast because you turned on CUDA graphs.

[Rich Stureborg @rstureborg]

The repo summarizes some of this but here are 2 of the more important problems they made progress on:

1. Quasi-Riemann hypothesis: This gives much stronger guarantees on how irregularly primes can be distributed, by restricting where zeros of the zeta function and related functions can occur. That said, this establishes a boundary at 7/8. the full Riemann hypothesis requires 1/2 and remains unsolved.

2. Unique Games Conjecture: New limits on how well polynomial-time algorithms can approximate optimization problems like Max-Cut and Vertex Cover. The caveat on this one is that interpreting it still assumes P /neq NP. Doesn't resolve P = NP.

[OpenAI @OpenAI]

We’re releasing a broad range of new mathematical results produced by an internal frontier model.

We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results.

https://t.co/7N6TPlft1P

[Rich Stureborg @rstureborg]

@TrulyWande44156 @claudeai ouch 😬

[Rich Stureborg @rstureborg]

@igcorreia @claudeai 😆😆

[Sam @sam_builds_ai]

When is a gift not a gift?

Anthropic’s startups program looks like a generous gift with free Team seats, API credits, and access to the Anthropic team.

Rich Stureborg read the fine-print you agree to by applying and found some pretty damning inclusions.

1) Anthropic can develop competitive tech.
2) Personnel can reuse general ideas and know-how retained in memory.
3) Safety reviews can override ZDR.
4) The addendum can change without notice.

That is the Trojan horse, and like the city of Troy, you will be opening your own doors to it.

Its a similar pattern as the containment problem I wrote about yesterday, however this time the containment of your IP is hidden in the fine print, and explicitly called out as not being protected (at worst collected and used!)

[Rich Stureborg @rstureborg]

@Jimmybonkers_ @claudeai 🤓

[Rich Stureborg @rstureborg]

https://t.co/6FjsevTJQz caught it!

[JimmyBonkers @Jimmybonkers_]

@rstureborg @claudeai Gotta love when someone with a legal mindset looks at the terms that no one reads lmao. Good catch

[Rich Stureborg @rstureborg]

@expectcuriously @claudeai yikes! i don't see any way for you to remove your acceptance of terms. did you find one?

[Rich Stureborg @rstureborg]

@mmeding and data!

[Rich Stureborg @rstureborg]

@HarveyMushman10 yup