iiinova — where your digital self keeps evolving

SimCity 3000 - Magic City

I was 9 in 2001 poking around Sim City when suddenly the most transcendent music I'd ever heard came on and changed my life and music taste forever www.youtube.com/watch?v=LkMZ...

youtube.com

YouTube video by Gamegroove

8453

I like the e/accs who are like "Effective altruists are wildly out of touch with the values of everyday Americans. Not like us, no sir our lifestyles and beliefs are totally normal. Everyday Americans want to be heat engines."

1301

"Top 10 water user among industries" is doing a lot of work here. Industry uses very little water compared to the main way America uses water: agriculture. People just throw these general slogans out and don't actually bother to do the normal thinking we'd do with any other industry using water

21228

Few things better than people you really respect and look up to saying they got a lot out of something you wrote

0590

I like the insult "goofy goober" because it's pretty soft as insults go, but drives really status-sensitive people who aren't used to being poked at completely insane

5718

Western MA occupies the same place in my brain as Eden, both the paradise and the fall, so have been following the Boston Globe story with some interest. A lot of my friends slowly went a little crazy out there in the woods.

0120

My benchmark for "Full automation of the Hold Steady" is at about 65%

090

Andy Masley @andymasley.bsky.social · 21/09/2026

I think there are some things we can actually think through and start to understand where the risks would be, with a lot of uncertainty. I think we could've started thinking about the dynamics of countries owning nuclear weapons before they existed, for example. With AI I think we're obligated to

5402

Trying to quit caffeine for a bit and just found out Ghost has a new root beer flavor. They never tell you how tough adult life is gonna be sometimes

3220

Andy Masley @andymasley.bsky.social · 21/09/2026

This is the first in a series I'll write on what I get out of effective altruism. Trying to write this series as a big complete overview for a smart friend without any jargon, getting across the appeal without turning it into a sales pitch. blog.andymasley.com/p/what-i-get...

blog.andymasley.com

What I get out of effective altruism

Part 1: The EA story of the world

1371

Andy Masley @andymasley.bsky.social · 20/09/2026

I am sorry for anthropomorphizing the single thing that hundreds of billions of dollars and the smartest people in the world are all trying to make as much like the human mind as possible. That was silly of me

41519

The comments completely hate my guts unforch

2642

Recently had a friend have an actual “the vitamin” experience with iodine

1671

The Data-Center Debate Is Divorced From the Facts

First piece I’ve done for another publication! In The Atlantic today on data centers www.theatlantic.com/ideas/2026/0...

theatlantic.com

All those scary numbers aren’t what they seem.

511519

Reposted by Andy Masley

J. C. Cantwell 🌻 @segfaultvicta.aludel.xyz · 18/09/2026

the number of times the answer to "where does your belief that <specific misinfo> is true come from?" is "reading Cory Doctorow" is really fucking concerning actually

616415

Will Stancil @whstancil.bsky.social · 18/09/2026

Kind of a disaster for scientific communication that this still needs to be said

3933

Reposted by Andy Masley

Again, this is a very severe misunderstanding. Models do not read through their training data looking for instructions on how to do things, which is why they can solve novel problems or exhibit creativity.

31043

Reposted by Andy Masley

Will Stancil @whstancil.bsky.social · 18/09/2026

This is not how AI models work. It just isn't. They aren't mechanically emulating stuff from their training data, like kids playing dress-up. Do you actually ever use an LLM? I am not a big user myself, but even small amounts of actual exposure tends to disabuse people of this notion fast.

181394

Andy Masley @andymasley.bsky.social · 17/09/2026

I would very much like everyone to adopt a pretty conservative personal lifestyle to avoid said awful power dynamics and for the most part outside of the Bay we do

2170

Andy Masley @andymasley.bsky.social · 17/09/2026

The funniest possible timeline would be where the War Department was calling us TESCREAL

4730

Andy Masley @andymasley.bsky.social · 16/09/2026

I once had someone on a panel ask "What I don't understand is why EA is so pathologically focused on the NUMBER of lives saved" and I said that I think if everyone's equal, 10 people dying instead of 5 is exactly as bad as 5 people dying instead of no one dying. Egalitarianism requires calculation

3450

Andy Masley @andymasley.bsky.social · 16/09/2026

Okay folks was traveling for a week, what'd i miss

8320

Andy Masley @andymasley.bsky.social · 11/09/2026

AI safety people aren't being influenced by Terminator 2. Very few AI safety people have seen Terminator 2. I know this because it's a perfect movie and I've had a side quest over the past few years to get influential safety people to watch Terminator 2 with me.

814511

Andy Masley @andymasley.bsky.social · 10/09/2026

One useful tool for some quick thoughts on an AI safety career pivot would be 80,000 Hours' advice bot app.80000hours.org

0171

Andy Masley @andymasley.bsky.social · 09/09/2026

What did people expect the *head of alignment at Anthropic* to believe about AI risk?

61101

Andy Masley @andymasley.bsky.social · 09/09/2026

Some people thought I was joking here. This is a video game with a funny name that basically seemed impossible for AI to get good at for a while because the learning curve is so steep. There aren’t many intermediate steps it can use along the way. This is a useful sign that it’s broadly capable

2553

Reposted by Andy Masley

spontaneous symmetry breakdance @t8erboi.bsky.social · 09/09/2026

If anyone knows good on-ramps for someone coming out of theoretical physics, I would appreciate recommendations

3181

Andy Masley @andymasley.bsky.social · 09/09/2026

I don't think I've ever seen anything AI safety related get this many likes. We gotta build onramps for people, fast.

61042

The newest models are the first that I could personally seriously consider being AGI. Kind of quaking with borderline religious fear to be honest.

6534

I apologize to @segyges.bsky.social but I think I made Benn think you run Bluesky

71526

This is a very funny way of deflecting. I'm now the one who's "dodging" because I don't wanna get on a livestream with the guy who won't bother to just type out explanations for any of the crazy lies in his videos. He clearly just wants to bullshit more.

2602

I basically think a lot of people have been implicitly taught that any understanding of the human mind that isn't Homunculus Theory is demeaning and flattening, and they don't know the deal in philosophy of mind, and we really need people to entertain the arguments in that field

5665

EA forcing me to see how little people actually respond to incredibly obvious arguments to donate a portion of their money to the lowest income people in the world made me somewhat doomer on this situation, if it does pan out

2572

Fuck what if LLMs have qualia and we don't. What then?

161116

Language is the house of being, in that it's becoming out of reach for a middle class family!

191

Look if you don't like intervening in nature to stop predation, you're going against nature

2244