Steve Zekany

My Claude has informed me that ordering Pad Thai is the tutorial dish and now I am emotionally devastated.

Trying something new: I wrote up some thoughts on Spirit Airlines. Core idea: unbundled pricing is more honest (or should be), and our cultural mockery of Spirit reveals an uncomfortable tension about how Americans link consumer choices to dignity.

Claude saved me $65 this year on some random foreign tax credit but the real clutch move was when I lamented that a former employer didn’t send a W2 and Claude was like “oh just go to irs.gov and pull your wage transcript, it probably has enough for you to file”.

The most fulfilling response you can get from Claude

Born too late to explore the planet, too early to explore the galaxy, but just in time to train on RTS games since elementary school, an Ender’s Game situation for an entire generation.

+1, this is a really excellent piece.

@ohlennart Glad you’re back!

speaking of putting essays on the internet, i'm back on X. will tweet a bit more and also put out one or two essays in the next couple of weeks.

I think most people haven’t updated enough toward not being able to trust Reddit posts in a billion tokens/sec world.

@SwannMarcus89 I think most people haven’t updated enough towards not being able to trust Reddit posts in a billion tokens/second world.

I would love some evidence this guy is who he says he is and isn’t just making shit up on Reddit lmfao

imagine running claude code with opus 4.5 here

This seems like a good analogy. Knowledge work done by AI is unlikely to follow the exact process as performed by humans, and the process it does follow will probably look unsettlingly strange at first glance.

@dioscuri @NathanpmYoung Learning styles is another one. A diagram explaining a concept isn’t helpful because you’re a visual learner; it’s helpful because some information is more effectively communicated visually than with text!

@NathanpmYoung The worst one by far in terms of public understanding imho is “we only use 10% of our brain.” Regularly encounter this misconception even among smart college students.

Bears don’t sleep all winter. What’s another commonly known “fact” that isn’t true?

@S_OhEigeartaigh @NathanpmYoung Frozen vegetables have as much or more nutritional value as fresh vegetables (flash frozen at peak ripeness). Being outside in the cold does not make it more likely you will catch a cold (actually less if the alternative is gathering indoors)

Spinach has iron, but not nearly as much as people think (was overestimated at some point in the past). Vitamin C prevents scurvy, but has minimal protective effects and only slight recovery-speeding-up against colds and illnesses in general. (Damn you, Mr Pauling). Milk isn't a health food.

@johnvmcdonnell Merry Christmas to you and the family too!

Merry Christmas everyone!

@aarmlovi A good illustration of what is wrong with blue state governance right now: avoiding expensive power outages for your constituents is just an evil and underhanded plot to ensure you’re more qualified the next time you run for elected office!

The data center moral panic is cancelled, folks! We're now pivoting to second-guessing Gov Newsom's bold decision to cheaply restore one of California's biggest reliable sources of carbon-free electricity

the US auto industry is going to be eaten alive by china's

Apply for the 2026 Summer Fellowship, now with two tracks! Fellows join GovAI in London for 3 months, conducting a paid project with mentorship from leading experts in the field. Apply by 4 January, 2026. 🔗 Learn more: governance.ai/opportunities

A lot of proposals the AGI governance space centre on the idea that governments (or populations, or other notable entities) will be rational actors, and that cooperating or competing entities will have a sufficiently similar understanding of The Situation to be modelled as rational actors cooperating or competing. I think this is a risky assumption. At the moment, if you pick any country - say, the UK - how AGI company leaders, political leaders, and civil society see the state of the science and its likely future (and governance implications) differ widely. There are very different expectations even within e.g. government. And that's before you even bring other countries into the conversation. If progress towards AGI is sufficiently slow, or if the 'normal technology' view is correct, then that will hopefully allow enough time for relevant entities to converged on a shared understanding of the technology and its strategic stakes. However if progress towards AGI is remotely fast (say, <10 year timelines), I don't expect to see that. I expect a lot of confusion and fog of war. And if other actors make what look like aggressive moves, there will be a bewildering array of misunderstandings and misinterpretations. So while I think there's some value to the 'rational strategy' type analyses, I'd love to see a lot more analyses that at least acknowledge the chaos, confusion and fog of war we're likely to be dealing with - and try to account for them, if possible.

@constans It’s such a perfect example of how trolley problems exist in real life. The short-term pain of coordinating which novels to read and holding middle school students accountable is so small compared to the downstream effects.

There’s going to be a massive educational opportunity gap between upper middle class students and everyone else. Academically inclined parents won’t tolerate this at all, and everyone else will be at the mercy of these low standards and be permanently crippled