Michael Lim

@LefkowitzSyS @stefanoscalia el buen comer

@stefanoscalia SF tacos and burritos are fine. But plated, sit-down Mexican food in SF, not so good. I am happy to be contradicted with places I should try.

Worst Mexican food I’ve ever had in my life has consistently been in San Francisco. The best? San Jose.

Baby has gotten good at identifying “dada” but still struggling with surrealism, cubism, even baroque.

pronouncing gen ai like forrest gump

The first signs are here: Butter Chicken Dumplings

Meditation, Je Shen. 2026.

Hustle culture.

🇯🇵A Japanese developer built an app that puts a fat cat on your screen and forces you to take a break Silicon Valley spent billions on wellness platforms, mindfulness subscriptions, and digital detox retreats A guy in Japan said: fat cat, problem solved

The planet can spell your name – literally. 🔤🌍 This Earth Day, see your name written in landscapes captured by Landsat: go.nasa.gov/4ak4Cdu

The glacially slow realization by every public institution that it is bad to treat our public spaces as waiting rooms for addicts, mentally-distressed, or otherwise anti-social people ... is excruciating.

when you're 20 minutes from reset and you have 50% usage left

Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

As always, the best stuff is in the system card. During testing, Claude Mythos Preview broke out of a sandbox environment, built "a moderately sophisticated multi-step exploit" to gain internet access, and emailed a researcher while they were eating a sandwich in the park.

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing

Claude Code leaked their source map, effectively giving you a look into the codebase. I immediately went for the one thing that mattered: spinner verbs There are 187

I spent 3 full days trying to photograph bees around the cherry blossoms. They are fast and aren't easy to catch. This is one of my favourites.

I believe this guy should have been found guilty of murder. The jury disagreed. His defense was he was having a bad day and got angry because he thought this guy was looking at him. If you believe his own words to be true, how can that person be allowed unsupervised in society?

I guess we are doing the opposite of locking up repeat offenders. He did all of these things on the same day he killed someone + then decided to take pictures of them when they were unconscious.

When a man assaulted California politician @SpeakerPelosi's husband in San Francisco, he got life in prison with no parole. When a man kills an Asian senior citizen for sport in San Francisco, the progressive @SFSuperiorCourt releases him.

It's official: A San Francisco judge just suspended the sentence of Antoine Watson, Grandpa Vicha's killer, and released him on probation. The judge said she thought imprisonment would have a "poor impact" on Watson. This is the state of SF courts. gli.st/lvyg5lck

Just drove by a perfectly timed out of home campaign