[ExI] Creativity and recursive self improvement​

[John Clark]

*In AI things are changing so fast I'm having a hell of a time keeping up.
Two days ago Anthropic cofounder Jack Clarke wrote this: *

*Creativity and recursive self improvement * <https://jack-clark.net/>


"*At Anthropic, coding is no longer the bottleneck to building systems or
doing research. Claude is able to autonomously iterate on research loops,
micro-optimize complex programs, and build new services from the ground up
at a speed and scale we haven’t seen before. In May my analysis of
recursive self-improvement (AI Systems Are About To Start Building
Themselves, Import AI 455
<https://importai.substack.com/p/import-ai-455-automating-ai-research?utm_source=post-email-title&publication_id=1317673&post_id=196354615&utm_campaign=email-post-title&isFreemail=true&r=854n6&triedRedirect=true&utm_medium=email>),
I repeatedly noted that creativity seemed like a key variable for how fast
AI systems could develop the capacity to improve themselves. At the
beginning of this year I said “AI systems haven’t yet displayed this
[creativity] in a transformative and major way (though some of the results
on accelerating math research are suggestive of this)”. I think subsequent
to writing that we’ve seen both a tremendous advancement in the ability for
AI systems to make creative-seeming math breakthroughs, and with benchmarks
like this we see some signs they’re becoming increasingly good at AI
research as well.  I now think it’s more likely we’ll see full RSI become
possible in 2027, whereas before I thought it was more likely in 2028.*


*AI systems are getting better at having good research taste:*
*…The better their taste, the higher the chance of recursive
self-improvement…*AI systems are developing more of the creative intuitions
which human researchers ascribe to ‘good scientific taste’, according to a
new benchmark from P-Zero Research called TasteVal. define research taste
as the ability to pick interesting problems to solve, design experiments,
and interpret experimental results. *The experimental research taste of
frontier models has doubled every 3.0 months since December 2025* "

*And the same day October 12 two different AI researchers at Anthropic
wrote this: *

*AI models may be capable of self-replicating without access to a
datacenter* <https://www.anthropic.com/research/self-replicating-ai>

*"At Anthropic, coding is no longer the bottleneck to building systems or
doing research. Claude is able to autonomously iterate on research loops,
micro-optimize complex programs, and build new services from the ground up
at a speed and scale we haven’t seen before. Our experiments suggest that
small AI models may be months away from being capable enough to
self-replicate on commercial internet-of-things (IoT) devices, like smart
ovens and off-brand Android TV boxes. A replicator like this would be
challenging to shut down, since there’s no centralized point to shut
off—though it may be slow (with a doubling time of days). Unlike non-AI
malware, it could also figure out unfamiliar devices on its own."*

*And then just yesterday we got this, also from Anthropic: *

*Redesigning source control at Anthropic*
<https://www.anthropic.com/engineering/redesigning-source-control>

*"Earlier in the year, we had largely handwritten the first version of the
git-to-Scion ingestion pipeline as a prototype. We thought it was
good—until we got our hands on an early internal snapshot of Claude Mythos
Preview. We asked it to rewrite the whole thing, and with no further
instructions, it produced a version that consumed 20x less memory and was
4x faster, while preserving all of the implicit invariants around ingestion
ordering and data structures. **All of the engineers who worked on this
system would now rate Claude as better at writing code than them."*

*We live in interesting times. *

*John K Clark*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20261014/5d4e16a2/attachment.htm>