Yes there are some extremely novel approaches. Scott Aaronson highlights[1] some of ones he's interested in, and I'd highlight:

> Matrix multiplication in O(n9/4) time—a rational exponent for once (!), and via a completely different approach than was used for O(n2.373) and so forth

And in his wife's comments on USG (which she worked on):

> The UGC proof invents a completely new bizarre code with a noise test. It’s some crazy recursive construction. It’s not the long code, not the short code – some alien craziness

[1] https://scottaaronson.blog/?p=10169

Did you have prompts that were especially tuned for Haiku 4.5 or something?

I'm not aware of any benchmarks that measure the first "analyze XYZ geopolitical situation" but legal reasoning is very closely related to "explain the ramifications of XYZ law".

Legal Bench[1] measures legal reasoning. It's close to saturated (ie, there isn't a lot of room for improvement) but Fable scores 88% vs eg Opus 4.7 at 85%.

There probably isn't a lot of room for improvement on something like this - there is enough disagreement in legal reasoning to mean 100% is going to be impossible.

[1] https://www.vals.ai/benchmarks/legal_bench

https://openai.com/index/navier-stokes-solution/

The industrial revolution produced economic disruption.

More recently I remember there were typing pools at my dad's work and there was serious concerns about what jobs women would do because personal computers meant they weren't needed.

Turns out in both cases people were blinded by biases around future work looking like current work.

> Following an investigation, we have confirmed that Buckmaster’s Codex prompts over the two months preceding this announcement and paper on September 8, 2026, could not have influenced the system in any way, including through training.

https://openai.com/index/navier-stokes-solution/

I agree with the courts. I don't think learning from something is piracy in anyway.

Obviously though this is a very different issue to what the OP was claiming. In that case there is no legal argument at all that they could train on it and the argument is there about moral rights.

https://www.nature.com/articles/s41598-023-46392-6

There's quite a lot of progressive positive writing.

On the economic side the Australian Council of Trade Unions statement is about a positive future: https://www.actu.org.au/speeches-and-opinion/joint-statement...

I for one think understanding more about how the world operates is just about the highest calling possible.

> when AI discovers the cure to cancer or aging etc.

A guy I knew did this. It successfully shrunk cancer tumours in his dog: https://www.the-scientist.com/chatgpt-and-alphafold-help-des...

Graph theory (which the OpenAI math results had many proofs in) is directly applicable to cancer modelling and drug design.

But sure. Novelty crap.

I think we should stop the bad things.

We don't stop building cars because there are crashes, we build better safety systems.

I'm against the doomer narrative ("AI will kill us all") not against a clear eyed approach to making safe systems.

Improvement in this context means "better quality results".

You can use better quality models to do worse things with.

I'm not making any claim about second order effects like that.

I think AI - like all changes - will lead to some bad things. The internet did too!

But I don't think AI will kill us all.

Interestingly I'd note that the two outcomes you listed (power concentration and loss of control) are dimensionally opposites!

For me this just shows that the future contains such a vast array of possible outcomes that focus on the negatives completely missed the positive outcomes that future also holds.

The sub n log n result is astonishing: https://github.com/openai/math/blob/main/preprints/Integer-m...

Here's a great article 2019 on the quest to achieve the n log n boundary:

> Schönhage and Strassen’s ungainly n × log n × log(log n) method held on for 36 years. In 2007 Fürer beat it and the floodgates opened. Over the past decade, mathematicians have found successively faster multiplication algorithms, each of which has inched closer to n × log n, without quite reaching it. Then last month, Harvey and van der Hoeven got there.

and

> Harvey and van der Hoeven’s algorithm proves that multiplication can be done in n × log n steps. However, it doesn’t prove that there’s no faster way to do it. Establishing that this is the best possible approach is much more difficult. At the end of February, a team of computer scientists at Aarhus University posted a paper arguing (opens a new tab) that if another unproven conjecture is also true, this is indeed the fastest way multiplication can be done.

As far as I'm aware no one seriously believed sub n log n multiplication was possible. It just seemed such a logically sensible boundary it was taken as true-but-unproven.

https://www.quantamagazine.org/mathematicians-discover-the-p...

This seems great!

What makes you say that? What is an example of a domain where the improvement is small?

I can't think of any at all. Compare something as unverifiable as "Make good music". Models now are many times better than 3 years ago.