Lectus: Readings on Technology and Intelligence.

A space for chosen thoughts, not noise. Tracking real developments in AI and emerging tech through multiple lenses; technical, economic, existential.

[Denny Zhou @denny_zhou]

Slides for my lecture “LLM Reasoning” at Stanford CS 25: https://t.co/eApGUHyIDo

Key points:
1. Reasoning in LLMs simply means generating a sequence of intermediate tokens before producing the final answer. Whether this resembles human reasoning is irrelevant. The crucial insight is that transformer models can become nearly arbitrarily powerful by generating many intermediate tokens, without the need of scaling the model size (https://t.co/HO2seV6vVl).

2. Pretrained models, even without any fine-tuning, are capable of reasoning. The challenge is that reasoning-based outputs often don’t appear at the top of the output distribution, so standard greedy decoding fails to surface them (https://t.co/75h2QQzT9M)

3. Prompting techniques (e.g., chain-of-thought prompting or "let’s think step by step") and supervised finetuning were commonly used to elicit reasoning. Now, RL finetuning has emerged as the most powerful method. This trick was independently discovered by several labs. At Google, credit goes to Jonathan Lai on my team. Based on our theory ( see point 1), scaling RL should focus on generating long responses rather than something else.

4. LLM reasoning can be hugely improved by generating multiple responses and then aggregating them, rather than relying on a single response (https://t.co/BA5MUzg3PR).

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

The gut feeling of an experienced Product Manager is a valuable addition to data when making product decisions. Sometimes/many times even more valuable than the data-driven decision approach. @AndrewYNg "Synthesizing all these sources of data in a PM's gut helps us move faster."

[Andrew Ng @AndrewYNg]

The invention of modern writing instruments like the typewriter made writing easier, but they also led to the rise of writer’s block, where deciding what to write became the bottleneck. Similarly, the invention of agentic coding assistants has led to a new builder’s block, where the holdup is deciding what to build. I call this the Product Management Bottleneck.

Product management is the art and science of deciding what to build. Because highly agentic coding accelerates the writing of software to a given product specification, deciding what to build is the new bottleneck, especially in early-stage projects. As the teams I work with take advantage of agentic coders, I increasingly value product managers (PMs) who have very high user empathy and can make product decisions quickly, so the speed of product decision-making matches the speed of coding.

PMs with high user empathy can make decisions by gut and get them right a lot of the time. As new information comes in, they can keep refining their mental models of what users like or do not like — and thereby refine their gut — and keep making fast decisions of increasing quality.

Many tactics are available to get user feedback and other forms of data that shape our beliefs about users. They include conversations with a handful of users, focus groups, surveys, and A/B tests on scaled products. But to drive progress at GenAI speed, I find that synthesizing all these sources of data in a PM's gut helps us move faster.

Let me illustrate with an example. Recently, my team debated which of 4 features users would prefer. I had my instincts, but none of us were sure, so we surveyed about 1,000 users. The results contradicted my initial beliefs — I was wrong! So what was the right thing to do at this point?
- Option 1: Go by the survey and build what users told us clearly they prefer.
- Option 2: Examine the survey data in detail to see how it changes my beliefs about what users want. That is, refine my mental model of users. Then use my revised mental model to decide what to do.

Even though some would consider Option 1 the “data-driven” way to make decisions, I consider this an inferior approach for most projects. Surveys may be flawed. Further, taking time to run a survey before making a decision results in slow decision-making.

In contrast, using Option 2, the survey results give much more generalizable information that can help me shape not just this decision, but many others as well. And it lets me process this one piece of data alongside all the user conversations, surveys, market reports, and observations of user behavior when they’re engaging with our product to form a much fuller view on how to serve users. Ultimately, that mental model drives my product decisions.

Of course, this technique does not always scale. For example, with programmatic online advertising in which AI might try to optimize the number of clicks on ads shown, an automated system conducts far more experiments in parallel and gathers data on what users do and do not click on, to filter through a PM's mental model of users. When a system needs to make a huge number of decisions, such as what ads to show (or products to recommend) on a huge number of pages, PM review and human intuition do not scale.

But in products where a team is making a small number of critical decisions such as what key features to prioritize, I find that data — used to help build a good mental model of the user, which is then applied to make decisions very quickly — is still the best way to drive rapid progress and relieve the Product Management Bottleneck.

[Original text: https://t.co/1tulDs3k7U ]

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

Interesting read from @svlevine arguing that surrogate data (simulation, human videos) is like a 'spork' - trying to get the best of both worlds but often just being mediocre at both. Makes the case for why real-world robot data is irreplaceable. https://t.co/ISMOMDH4Ei

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

The most important insights emerge between competing narratives.

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

Hugely impressive achievement by the @OpenAI team.

[Alexander Wei @alexwei_]

1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

But on its way...

[OpenAI @OpenAI]

To be clear: We’re releasing GPT-5 soon, but the model we used at IMO is a separate experimental model. It uses new research techniques that will show up in future models—but we don't plan to release a model with this level of capability for many months.

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

Couldn't agree more. This applies to new technology too. The most valuable AI applications will be anchored in deep understanding of established fields, not just novelty for its own sake. Wisdom before wizardry.

[Prof. Feynman @ProfFeynman]

You're unlikely to discover something new without a lot of practice on old stuff.

[Ryan Holiday @RyanHoliday]

Be strict with yourself and tolerant of others, Marcus Aurelius says. Your standards are for you.

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

Law and policy must evolve. But to do so wisely, they must first acknowledge the limits of their own language.

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

A Right No One Can Enforce
In 2023, the EU finalized the world’s first comprehensive AI regulation. Among its many principles is a commitment to transparency and explainability — ideals that feel noble and necessary in the age of inscrutable neural networks.

[Lectus: Readings on Technology and Intelligence. @lectus_insights]

Conclusion: A Regulatory Reckoning
The illusion of explainability isn’t just a legal flaw—it’s a philosophical reckoning.

We are building systems that exceed the frameworks we’ve relied on for centuries: responsibility, causality, rationality.