Co-founded Bullet Point Network, which runs the family office for a Goldman Sachs pre-IPO partner & helps investors to Quantify their Stories in Scenarios
San Mateo, CA
Joined April 2015
- Tweets2.2K
- Following466
- Followers477
- Likes1.7K
Yes, gasp, this could be the time for product managers to shine...
Product management for AI agents is easily the wildest form of product management in history.
Typical product management is trying to figure out how to design interfaces and software for people to interact with deterministic systems. The user generally knows all the context to do their work successfully, so it’s generally a matter of nailing the underlying business logic and surrounding UX.
But with AI agents, the user you care about most is the agent, and they don’t know anything by default. They’ll happily run in any direction to perform the task, often without success.
So as a PM (or engineer) you basically spend your time trying to reverse engineer “what would a human need as context to perform this task”, and then figure out how to design systems to get the agent that data in the right sequence, with the right tools, and instructions.
Some of these systems are entire invisible to the human user, but part of the craft is equally how the end-user will interact with the agent to supply this context. Then, it’s often unending trial and error working to eke out incremental points of quality at each stage.
This is especially why people with deep domain expertise, or those that can acquire it quickly, will do extremely well building AI agents. The ability to anticipate the context that the agent would need to be successful is a huge determining factor in how effective the agent will be.
This partly explains why coding agents have worked so well out of the gate; because its builders deeply understand the domain that they’re working to automate. But clearly we’re going to quickly see this same outcome across every field - legal, healthcare, finance, etc. - as context engineering and a new crop of product managers emerge.
Peter Moore retweeted
That’s 10% of US GDP. And over half of all US capital spending (it’s mostly being spent here). Which is a lot.
$NVDA sees $3-4 trillion in datacenter capex before 2030.
That is an incredible number....maybe a tad optimistic.
Changes in expectations for other things in other sectors are rippling from this change in expectations described this week by Sequoia's David Cahn: "Last year, the debate over AGI’s imminent arrival was raging . . . Today, the narrative has changed . . . the idea is that we can grind out serious progress from today’s models without considerable jumps in baseline intelligence . . . But even as this new narrative has become dominant, few people have come to terms with its full implications. For example, if progress is likely to come from domain specific applications of this technology, rather than exponential compute scaling, do we still need to build multi-million GPU clusters and add 90 GW of energy to the US power grid over the next 5-10 years?"
Yes! And as you accumulate info & interest about related topics, you apply the new book to build on top of the ideas that you've been building about these topics.
Ezra Klein: Part of what's happening when you spend seven hours reading a book is you spend seven hours with your mind on a given topic. But the idea that ChatGPT can summarize it for you is nonsense.
The point is that books don't just give you information. They give you a container to think about a narrowly defined scope of ideas.
Downloading information is obviously part of why you read books. But the other part is that books let you ruminate on a topic with a level of depth that's hard to achieve on your own.
@ezraklein
Q: "If natural language becomes the new code, then for new startups, what is their edge? Is it distribution? Is it speed to execution? Is it the idea?"
Google DeepMind CEO Demis Hassibis: "Well, I think it could be all 3 of those things. It could also be some deep vertical integration with maybe some specialist data or some specialist knowledge. I think there’s a lot of room for hybrid systems in the next few years like AlphaFold. You know, AlphaFold obviously has deep learning in it. But it’s not just that. It also has some biochemistry and physical chemistry constraints built into it. So I actually would like to see a flourishing of those kinds of hybrid systems."
Peter Moore retweeted
8K with SEC filed April 30. We are officially a kleptocracy.
An international trucking logistics firm is buying as much as $20 million worth of President Donald Trump’s crypto coins to influence the administration’s trade policy huffpost.com/entry/trump-coi…
In 4Q2025, skilled humans were still better than the AIs at judging the odds for future events, although by less than before: metaculus.com/notebooks/3529…
As I have for years, I continue to think that Ray Dalio's historical framework can help us anticipate the big political and economic changes happening right now and still to come. For those still not ready to read his 2021 book (which he started publishing on LinkedIn in 2019), here's his video summary: youtube.com/watch?v=xguam0TK…
"That’s why it is not just coincidence that the US, with its deep, flexible and well-governed financial markets has manufacturing shares of GDP well below the global average, unlike economies such as China’s with persistent surpluses, who have manufacturing shares well above the global average. Industrial policies aimed at restructuring more highly controlled domestic economies also in effect restructure the economies of their more open trade partners . . . However if the world is unable to come to such an agreement, the US is justified in acting unilaterally to reverse its role in accommodating policy distortions abroad, as it is doing now. The most effective way is likely to be by imposing controls on the US capital account that limit the ability of surplus countries to balance their surpluses by acquiring US assets. While this may at first seem to go against current US policy under Trump, who wants to increase foreign direct investment, if done correctly capital controls would in fact have little effect on direct investment. A less effective way is through controls on the US trade account, with bilateral tariffs an especially clumsy way of addressing the root causes of trade imbalances." ft.com/content/a3fb6b82-4a51…
As a consequence of NCAA football & basketball revenue now being paid to football & basketball players, the NCAA's #1 men's swim team had to cut half its athletes: thedailytexan.com/2025/04/14…
Maybe this can work: "The idea is to extract commitments from U.S. trading partners to isolate China’s economy in exchange for reductions in trade and tariff barriers imposed by the White House. U.S. officials plan to use negotiations with more than 70 nations to ask them to disallow China to ship goods through their countries, prevent Chinese firms from locating in their territories to avoid U.S. tariffs, and not absorb China’s cheap industrial goods into their economies." wsj.com/politics/policy/u-s-…
“Trump is taking the cudgel to elements of the U.S. institutional framework that have long been seen as free from direct political interference. This could have serious long-term ramifications for the value and broad use of the dollar in global markets,” said Eswar Prasad, an economist at Cornell University. wsj.com/economy/central-bank…
Peter Moore retweeted
I’m playing everyone’s favorite party game: guess the Trump administration’s strategy. Like most of you, I have no inside information. But that’s what makes it fun. Here’s my best theory.
Some assumptions first:
1) They’re not crazy. There is a strategy. It doesn’t align with conventional economic principles. But there’s something they’re playing toward.
2) They’re not stupid. I know enough of the players involved to know they’re not idiots. They may be making what will turn out to be accurately predicted to be horrible decisions, but they do have a plan and it’s coherent. As part of this, while I think they think they can increase U.S. manufacturing, they don’t really believe they can bring everything on-shore.
3) They’re intentionally being opaque as to what the real plan is. Trump fashions himself a negotiator. Holding his cards close to his vest, even lying about what cards he has, is part of the game.
4) They’re not just in it for themselves. I get that this has become non-conventional wisdom, but I am going to assume for this that the goal isn’t merely grift. Even if you believe it is, suggest that’s an easy out to thinking through what may be more complex motivations.
5) China is the real enemy. China has done some things in the last two years that have made even the China doves in the last administration into hawks. Again, I know there’s lots of media saying Trump wants to kiss up to Xi. But, having spent enough time with enough folks in this and the last administration, they really worry about China morning, afternoon, and all night.
So with that context, I posit the strategy is entirely: destabilize and ultimately decapitate China.
If that’s right, you could just impose tariffs on China. But China has lots of export markets, so goods will just flow out through those. What if instead you impose tariffs on everyone?
While China and the U.S. have similar GDPs, China is much more dependent on exports. That means while the countries of the world like cheap Chinese exports, they don’t depend on the Chinese market for their exports (yet). China has actually been on a nationalistic spree recently so foreign brands are more out of favor, meaning they’re an even less interesting market to sell to.
The U.S., on the other hand, is the world largest buying market. We are the consumers to the world. If we stop buying, everyone suffers. That means nearly everyone needs to come to the table with the U.S. if there are universal tariffs.
The U.S. is also unique in that it is among the only countries that doesn’t need to import anything. Don’t get me wrong, we want to import iPhones and PlayStations and French wine and German cars. But we don’t need oil or food or water or most the other raw materials to make sure people stay alive.
So the U.S. will hurt under a high tariff regime but won’t collapse. Some manufacturing will move back on shore. But the biggest thing is every country needs to negotiate with the U.S.
What does the U.S. ask for? I’m sure a bunch of nits with every country. But what if the big ask is: it’s us or them. You either trade with the United States, or you trade with China, but trading with both isn’t acceptable anymore.
For some countries — Vietnam, Indonesia, Philippines — the promise is to be the next China, but this time under more careful rules dictated by the U.S.
For most the rest of the world, they’re already massive net importers from China. Forced to make a choice between selling to China or selling to the U.S. I’d guess most will pick the U.S. German automakers are terrified of BYD. And the Italians and French haven’t proven they can sell wine or cheese to China at any real volume.
What’s China’s response? It’s tricky because they’ve preached self-sufficiency and internal focus. But if the whole world order suddenly aligns against them, what do they do?
I have no idea if this is the Trump administration’s plan. But it’s the only thing I’ve come up with that passes the sniff test.
Peter Moore retweeted
Oh my goodness. GPT-o1 got a perfect score on my @CarnegieMellon undergraduate #math exam, taking less than a minute to solve each problem. I freshly design non-standard problems for all of my exams, and they are open-book, open-notes. (Problems included below, with links to GPT-o1's answers.)
While eating Pie in the afternoon, I showed the exam to one of our math Ph.D. students (a former International Mathematical Olympiad Gold Medalist from Belarus), and he said "Hmm. Non-Trivial. Good." Our undergraduate students are also very good. This exam was not easy for them, as the score distribution shows.
Today is the 2-year anniversary of the public release of GPT-4. Two years ago, it caught my eye because it exhibited sparks of insight, similar to what I would see when I talked to clever kids who learned quickly. That gave me the instinct and urgency to start warning people. Today's observation of GPT-o1 being able to ace my hard college exam, makes me feel like we're close to the tipping point of being able to do moderately-non-routine technical jobs. I was impressed by every student in my class who got a perfect score. The fastest such person took 30 minutes.
And GPT-o1 only costs $60 per million words output, which means that each problem cost about 5 cents to solve. A total of around 25 cents, for work that most people can't complete in 1 hour.
Problem 1: Consider the recurrence a_n = a_{n-1} + a_{n-2}, with the first initial condition being a_0 = 1. Find all real number values for the second initial condition a_1 such that lim_{n \rightarrow \infty} a_n = 0.
chatgpt.com/share/67d4d4bf-1…
Problem 2: Find coefficients such that the sequence a_n = n \sqrt{2} + 2^n \pi satisfies the following recurrence, for some initial conditions. You don't need to find the initial conditions.
a_k = ___ a_{k-1} + ___ a_{k-2} + ___ a_{k-3}
chatgpt.com/share/67d4d4fe-6…
Problem 3: Fill in the blanks. The middle entry of the result of:
[[0, 0, 1], [1, 0, 0], [2, 3, 4]]^n [[5], [6], [7]]
is the term a_n of this recurrence:
a_k = ___ a_{k-1} + ___ a_{k-2} + ___ a_{k-3}
a_0 = ___
a_1 = ___
a_2 = ___
chatgpt.com/share/67d4d61d-7…
Problem 4: Consider the recurrence with initial condition a_0 = 1, where for each n \in {1, 2, 3, ...}:
a_n = \sum_{k=0}^{n-1} a_k
Find the generating function f(z) = a_0 + a_1 z + a_2 z^2 + ..., which looks something like f(z) = (1-2z)/(1-z)
chatgpt.com/share/67d4d63c-1…
Problem 5: Prove that the coefficient of x^{2025} in (x + x^2)^0 + (x + x^2)^1 + ... + (x + x^2)^{2025} is a Fibonacci number
chatgpt.com/share/67d4d657-4…
My main work nowadays is to build and scale up a community of people (through education) to face the challenges of the AI age together. I thought I had more years. Now we have to move faster. – at Moon, PA
Peter Moore retweeted
Performing just 10 bodyweight squats every 45 minutes during an 8.5-hour workday outperforms a single 30-minute walk in regulating blood glucose
Why? Short bursts of intense muscle engagement trigger a stronger metabolic response than steady, low-intensity movement
The key player: Lactate—a powerful signaling molecule
It drives GLUT4 transporters to the cell surface, pulling glucose from the bloodstream into muscles
And the best part? This effect can last up to 48 hours
Great summary of the core tradeoff reality behind the past month's controversies.
The story of the penny shows how futile it is to rely on Congress for spending cuts. That leaves two terrible options: either the president disregards existing laws and the Constitution or gives up on cutting spending.
Nixon’s admin proposed eliminating the penny in 1976. However, the Constitution is very clear: “The Congress shall have power ... to coin money [and] regulate the value thereof.” Congress has to act.
Reagan, Bush and Obama all favored eliminating the penny to save money but waited on Congressional action that never came. Bills were introduced in '89, '01, '06, '11, and '13. They never got any momentum. A NYT story last year had a good account of the hurdles the effort has faced.
The legislative branch is the most democratic of the 3. It is closest to the people. Members of Congress also represent subsets of the population. As such, any spending cut affects at least some members' districts while the benefits (lower govt debt) are more vague. Voters haven’t prioritized deficit reduction, so ignoring spending cuts— even on wasteful things like the penny—is the path of least resistance.
And here we are: Wait for Congress, and nothing happens. Act unilaterally, and break the system of checks and balances.
I do think we need to prioritize spending cuts. I do not think the President should ignore the Constitution or existing laws. Congress must do the work, though that's easier said than done.
Peter Moore retweeted
I’m guessing this is the 1st 40% that Don just found under the White House couch cushions…
