![[YOU] on AI · Daily](https://particle.news/cdn-cgi/image/format=auto,width=208/https://cdn.particle.pro/url/media/b52b3f28-ae4e-455d-a2fe-4753fa1d6a8c/30eed5dfd8ce68124aa07b23f281cef8c044caef30738a6810ef2db44cf25a03)
Oct 17, 2026 · 19 min · 10 segments
On Saturday, October 17, 2026, the AI world confronted a day of historic convergence: Claude Mythos 5 redesigned quantum circuits by up to 17 times fewer gates, DeepSeek dropped a massive open-source…
Edo SegalHostThanks, Ido.
So, quantum computing has a resource problem.
The algorithms researchers most want to run demand more operations and more qubits than today's machines can comfortably handle.
Designing better circuits is painstaking expert work, and it has historically been done entirely by human specialists.
That changed this week.
In one case involving elliptic curve point addition, Claude's design used 17 times fewer gates than the earlier approach.
The method here matters as much as the result.
The team gave Claude plain language specifications and an automated evaluator that checked correctness and scored each design.
Claude worked in a closed loop, writing circuits as Python, receiving feedback, and improving without human input between rounds.
In one 53-hour autonomous run, it developed an entirely new reversible modular inversion method.
The paper, Archive 2610.22367, is from Anthropics STEM Fellows Program alongside Harvard and Cuera Computing.
The finding suggests that automated circuit design could accelerate progress toward practical quantum computing.
It's one of those results where you read it twice to make sure you're not misreading it.
Thanks, Ido.
So, quantum computing has a resource problem.
The algorithms researchers most want to run demand more operations and more qubits than today's machines can comfortably handle.
Designing better circuits is painstaking expert work, and it has historically been done entirely by human specialists.
That changed this week.
In one case involving elliptic curve point addition, Claude's design used 17 times fewer gates than the earlier approach.
The method here matters as much as the result.
The team gave Claude plain language specifications and an automated evaluator that checked correctness and scored each design.
Claude worked in a closed loop, writing circuits as Python, receiving feedback, and improving without human input between rounds.
In one 53-hour autonomous run, it developed an entirely new reversible modular inversion method.
The paper, Archive 2610.22367, is from Anthropics STEM Fellows Program alongside Harvard and Cuera Computing.
The finding suggests that automated circuit design could accelerate progress toward practical quantum computing.
It's one of those results where you read it twice to make sure you're not misreading it.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.