Top 25 AI Articles on Substack

Latest AI Articles




AI brain control, AI designs antibodies, realtime video games, Ernie, Pangu, Ovis-U1, Centaur, XVerse

Welcome to the AI Search newsletter. Here are the top highlights in AI this week.
Chai-2 is a revolutionary AI model that can design antibodies from scratch with unprecedented success rates, making it a game-changer for disease treatment and research. It achieves double-digit hit rates in de novo antibody design, outperforming previous methods by orders of magnitude. This breakthrough enables rapid and precise molecular engineering, …
AI Search








Why the CFO Tech Stack is broken

Here's how you can integrate AI in finance
The traditional CFO tech stack is siloed and inefficient
finstory AI and AI CFO Office ∙ 10 LIKES
Anthony Hasbun's avatar
Anthony Hasbun
Magnificent article. Forward-thinkers and iterative mindsets are not common among the accountants and corporate finance professionals out there. I like how some of these points sound similar to product management.
I think the newer generation will be ready to fully leverage AI and for us more senior, we can’t forget to constantly refine our skill set.


What will the IMO tell us about AI math capabilities?

Most discussion about AI and the IMO focuses on gold medals, but that's not the thing to pay most attention to.
This year’s International Mathematical Olympiad (IMO) will take place on July 15th and 16th in Sunshine Coast, Australia. It is the pinnacle of high school math competitions. Much like the Olympic Games, the stakes are national pride and personal glory.
Greg Burnham ∙ 10 LIKES
Parker Whitfill's avatar
Parker Whitfill
FYI there is a typo where "It’s entirely possible that we see no improvement. It’s perfectly plausible that AlphaProof gets 2-4 problems and LLMs get 1-2 problems. This would be consistent with no progress over prior capabilities, or could just be due to the problems being unusually hard for AI systems" is repeated twice.
Steeven's avatar
Steeven
I guess what’s the point of the IMO if you’re not exactly looking at the IMO problem set but only if the LLM can solve the problem in a particular way? I think it’s also possible that mere grinding might be enough to solve open problems and even if the LLM isn’t creative in a way that would be particularly impressive, it might not be a dead end for the LLM to simply get faster at grinding


AI Search
Jun 29

AI decodes DNA, 3D model animator, Hunyuan-GameCraft, Flux Kontext dev, VeriGen, Gemini CLI

Welcome to the AI Search newsletter. Here are the top highlights in AI this week.
AnimateAnyMesh is a new AI framework that can create high-quality animations for 3D models of any shape or size in just a few seconds. It uses a special technique to compress and reconstruct 3D models, allowing it to generate animations that are both realistic and diverse. This technology has the potential to make creating 4D content more accessible and…
AI Search

Nuclear Non-Proliferation Is the Wrong Framework for AI Governance

Placing AI in a nuclear framework inflates expectations and distracts from practical, sector-specific governance.
Michael C. Horowitz and Lauren A. Kahn — June 27, 2025
AI Frontiers ∙ 3 LIKES
Bruce Nappi's avatar
Bruce Nappi
As one of the team members who contributed to establishing the "non-proliferation" agreements in the 1970s and 80s, I agree with the higher-level objectives of this article. What's clearly missing, however, is the perspective of someone who actually knows what nuclear weapons are, how they're used, and what the drivers of the treaties really were. For example, in the section titled," How AI Differs from Nuclear Technology", it uses the criterion of similar social "potential" for both, and the threat to "human survival" they have. It then "claims" the comparison "breaks down for three reasons: AI’s far-reaching scope, its lack of excludable inputs, and its graduated strategic impact." The comparison, in fact, easily breaks down for very different, simple comparisons.
At the time the non-proliferation treaties were established, the nuclear "landscape" was very precisely defined. The U.S. had a stockpile of about 20,000 warheads. The U.S.S.R had about 70,000. We knew exactly how they were built. They knew how ours were built. We both knew generally: where both were stored, the command and control each had to launch them, and how they would be delivered. We both also knew, with high reliability, that no other country either had or was capable of rapidly building such stockpiles. This "landscape" is entirely absent from the AI universe.
The "less excludable" claim is also misleading because it is too focused on the exotic parts like "plutonium and uranium", which it views as able to be "restricted or controlled". That factor has long been lost. Right now, NINE countries have nuclear weapons. They have been sold or provided to others. Just a single use will radically change world politics. Beyond that, the creation of "dirty" nuclear bombs is within reach of any country that has nuclear power reactors. Past governments knew these issues. Each required a different type of overview and control "agency".
The third element, that nuclear weapons have "binary ... strategic" influence, overlooks the factors throughout human history that have led to the dominance of "nations" within their practical influence range. Egypt, Persia and Rome each could claim "world dominance" for their "time". But that was not based on some super weapon. It was a result of the level of "transportation technology". The "mast and sail" technology of the "tall ships" of Portugal, Spain, France and England was as impactful as their gun powder weapons.

AI Gaze
Jul 5

Agentic AIs —Are we creating autonomous Employees?

The AI Gaze
The integration of Agentic AI into our workflows and daily routines raises a profound question: Are we creating autonomous employees? As critical as this question is, several opinions have been given on this subject by Top AI professionals, which are quite opinionated. Let me quote some of them below:
The AI gaze ∙ 14 LIKES
Balogun David's avatar
Balogun David
I learnt a lot. Thanks for this 🙏
Razjaj's avatar
Razjaj
Amazing

Epoch AI
Jun 20

AI and explosive growth redux

Two updates from our integrated assessment model of AI automation
The debate around the macroeconomic effects of AI has shown no sign of convergence.
Andrei Potlogea and Anson Ho ∙ 1 LIKES

HUX AI
Jul 11

What is the HUX AI Research Internship?

Everything you need to know about our Summer 2025 Research Internship program - Q&A
We’ve designed the HUX AI Research Internship as a flexible, project-driven opportunity for students and early-career professionals who want hands-on experience at the intersection of responsible AI, data governance, and policy. It’s an internship, but with a strong focus on impact.
Merve Ayyuce KIZRAK and HUX AI ∙ 4 LIKES

AI & the retraining challenge

Historically, US government programmes haven’t helped much.
In this essay, Julian Jacobs writes about the history of US public worker retraining programmes, their efficacy, and how they might fare as AI diffuses throughout the economy.
AI Policy Perspectives and Julian Jacobs ∙ 14 LIKES
Logan Thorneloe's avatar
Logan Thorneloe
This was a super interesting read. Thanks Julian!



How to think with AI

The biggest opportunity for strategic finance is here.
In the demanding world of finance, every minute, every decision, and every ounce of focus counts.
AI CFO Office ∙ 17 LIKES