iiinova — where your digital self keeps evolving

Reposted by David R. MacIver

Antithesis @antithesis.com · 12/08/2026

What happens when your open source library changes millions of lives? To kick off season 3 of the Bug Bash podcast, we sit down with @drmaciver.bsky.social, Zac Hatfield-Dodds, and Liam DeVoe — three generations of Hypothesis maintainers, the PBT library used by ~5% of all Python devs.

141

David R. MacIver @drmaciver.bsky.social · 03/07/2026

(The port isn't perfect, but it's remarkably good and it's a hell of a lot better than having to run Python in a subprocess even with its flaws)

030

David R. MacIver @drmaciver.bsky.social · 03/07/2026

Yeah, sorry, we're a bit behind on updating the docs on hegel.dev. Eliminating the requirement was always on the agenda but it was a bit "lol maybe when LLMs get good enough to port Hypothesis to rust for us" and uh... turns out that happened already.

120

David R. MacIver @drmaciver.bsky.social · 18/05/2026

I can imagine there being ways to resolve this, especially with private markets in high trust environments, but I'm not optimistic.

010

David R. MacIver @drmaciver.bsky.social · 18/05/2026

(see e.g. Tetlock's work on superforecasting) Sadly the reality seems to be that that skill is outcompeted by people who have insider information or manipulate the market.

110

David R. MacIver @drmaciver.bsky.social · 18/05/2026

The steelman is roughly: There is a genuine skill and effort required for synthesising accurate probabilistic predictions about the future from publicly available information, those probabilities have social value, and a prediction market rewards skilled people for putting in effort to produce that.

210

David R. MacIver @drmaciver.bsky.social · 24/03/2026

So I joined Antithesis a little while back. When I did, I pitched them on a crazy idea. Antithesis... Hypothesis... Hegel! And here we are. Hegel is our attempt to do Hypothesis-grade PBT everywhere. Today: Rust. Tomorrow: The world! (Well. In a week or two: Go) antithesis.com/blog/2026/he...

antithesis.com

Hypothesis, Antithesis, synthesis

Introducing Hegel, our new family of property-based testing libraries.

150

David R. MacIver @drmaciver.bsky.social · 07/02/2026

I've largely got out of the habit of posting on social media to make number go up, but it feels somewhere between appropriate and darkly ironic to do that for my review of @add-hawk.bsky.social's new book, "The Score". drmaciver.substack.com/p/how-to-be-...

020

David R. MacIver @drmaciver.bsky.social · 04/02/2026

I'm definitely going to be shoving copies into several people's hands.

020

David R. MacIver @drmaciver.bsky.social · 04/02/2026

Equally awkwardly I've logged onto Bluesky for the first time in months to tell you that the book is great. It's like... 80% familiar ground for me, as I've read a lot of the source material (both yours and others), but even the 80% is clarifying a lot of details.

110

David R. MacIver @drmaciver.bsky.social · 16/10/2025

Yeah! I think I pass the tests with toilets (at least UK ones. I know that US ones do different things with water that I'm not totally clear on the details of, and there's a whole separate category of mains pressure toilets that work differently again) but lots of things I don't.

010

Reposted by David R. MacIver

Ben Lipkin @benlipkin.bsky.social · 13/05/2025

Many LM applications may be formulated as text generation conditional on some (Boolean) constraint. Generate a… - Python program that passes a test suite. - PDDL plan that satisfies a goal. - CoT trajectory that yields a positive reward. The list goes on… How can we efficiently satisfy these? 🧵👇

2136

David R. MacIver @drmaciver.bsky.social · 10/03/2025

I do not feel more zen now Dave.

010

David R. MacIver @drmaciver.bsky.social · 10/03/2025

Post a photo you took with no context to bring some zen to the timeline

140

David R. MacIver @drmaciver.bsky.social · 07/03/2025

(I don't think godlike intelligence is literally impossible, but I do think my probability that we ever get there is significantly under 50%)

000

David R. MacIver @drmaciver.bsky.social · 07/03/2025

But if "superintelligence" means something closer to godlike intelligence, I might be at never. I guess my working definition is "clearly superhuman on some axis, or peak human on multiple axes you'd not normally find together, and not clearly subhuman in easy to notice ways"

110

David R. MacIver @drmaciver.bsky.social · 07/03/2025

I'm honestly not sure how to answer this, and I think a lot of that is weighted on my uncertainty on what "unambiguous superintelligence" means. Erring on the side of stricter criteria I think 2050ish is probably the closest I'd confidently put at >= 50% probability.

100

Reposted by David R. MacIver

Laurence Tratt @ltratt.bsky.social · 06/03/2025

Shrinkray has very sensible rules that apply well to languages it has no specific knowledge of -- we make a *lot* of use of this and it's brilliant! -- and it scales near-linearly to the number of cores. If you need a reducer, I highly recommend it! Thanks @drmaciver.bsky.social!

191

David R. MacIver @drmaciver.bsky.social · 08/02/2025

I wrote some notes on how LLMs work, aimed at non-programmers, that you might find interesting. notebook.drmaciver.com/posts/2025-0... Feedback, corrections, and follow-on questions very welcome.

notebook.drmaciver.com

How do LLMs work?

291

David R. MacIver @drmaciver.bsky.social · 01/02/2025

Started reading this yesterday. It's very good so far, thanks!

000

David R. MacIver @drmaciver.bsky.social · 22/01/2025

Reminded how my first ever conference talk was a lightning talk about how microservices were bullshit. IIRC you were involved in pushing me into doing it, but I may be conflating events.

110

David R. MacIver @drmaciver.bsky.social · 30/12/2024

Mostly that bluesky is in the odd position of being both decentralised identity and also having a canonical identity that most people will use. ciphergoth.bsky.social is much more plausibly legitimate than ciphergoth.someother.domain

110

David R. MacIver @drmaciver.bsky.social · 30/12/2024

That's good, although it sortof feels like this is a nontrivial threat model for scams even without renaming. e.g. this wouldn't have been much better if @ciphergoth.org had always been at the .org rather than previously having the .bsky.social address.

110

David R. MacIver @drmaciver.bsky.social · 30/12/2024

Looks like the account has been suspended at least, but yikes.

110

David R. MacIver @drmaciver.bsky.social · 28/12/2024

Possibly part of the difference is that my balance just isn't very good, so I end up relying a lot more on visual cues to keep me upright, and if I were better at balance I wouldn't need that compensatory strategy.

000

David R. MacIver @drmaciver.bsky.social · 18/12/2024

I genuinely liked the righteous mind and still find bits of it useful. I also straightforwardly disbelieve most of its empirical claims.

010

David R. MacIver @drmaciver.bsky.social · 18/12/2024

My mental categorisation of Haidt is "often interesting, rarely correct"

110

David R. MacIver @drmaciver.bsky.social · 18/12/2024

Every time you wonder how something in Hypothesis works, you can usually reverse engineer it from the following design logic: 1. What's the obvious, natural, way to do it. 2. What are the problems with that? 3. Can those be fixed? Not without doing something ridiculous and unreasonable. 4. Do that.

040

David R. MacIver @drmaciver.bsky.social · 17/12/2024

Right, it's not that replies are excluded, it's that whether or not it's a reply is irrelevant, it's only explicit mentions that count.

010

David R. MacIver @drmaciver.bsky.social · 17/12/2024

@defenderofbasic.bsky.social I think that (possibly in a few minutes) this message is going to show up in your search.

020

David R. MacIver @drmaciver.bsky.social · 17/12/2024

Actually I have a theory. It's a slightly ridiculous one but I think it's possibly right. I think replies don't count. I think only explicit mentions do. I'm going to test this shortly one second.

210

David R. MacIver @drmaciver.bsky.social · 17/12/2024

Yeah on closer inspection, you're right, sorry. It's missing a lot of posts.

120

David R. MacIver @drmaciver.bsky.social · 17/12/2024

Although entertainingly the link *doesn't* because it strips the query parameters. Sigh.

010

David R. MacIver @drmaciver.bsky.social · 17/12/2024

Seems to work for me bsky.app/search?q=to%...

bsky.app

200

David R. MacIver @drmaciver.bsky.social · 16/12/2024

Yeah, I was wondering! Face up is easy, even without aikido. Face down is hard.

000

David R. MacIver @drmaciver.bsky.social · 16/12/2024

(The circumstances under which homemade ice cream makes sense, for the interested: 1) You must have space for a dedicated ice cream maker. None of this in-freezer bullshit. Ice cream maker or gtfo. 2) You must be making flavours with no store equivalent, or have garden ingredients you want to use.)

020

David R. MacIver @drmaciver.bsky.social · 16/12/2024

This is probably a skill issue, but I've always found this with hummus. Making ice at home is a mug's game. Making ice cream at home... there are circumstances under which this makes sense, but 99+% of people aren't in those circumstances.

110

David R. MacIver @drmaciver.bsky.social · 16/12/2024

The latest one, which you should be more careful with because I injured myself doing it (people who are less middle-aged-and-hypermobile are more likely to be OK): Can you get up to standing from lying face down on the floor, without rolling to the side or using your hands?

220

David R. MacIver @drmaciver.bsky.social · 16/12/2024

Something I very much enjoy on discord is when one of us memes the others into trying ridiculous exercises. e.g. a while back we had a conversation about whether it was easy to stand on one leg with your eyes closed (it's not) and we had like a dozen people going "Surely it is..." and trying it.

260

David R. MacIver @drmaciver.bsky.social · 15/12/2024

I think partly this is just that we're using the word "conflict" somewhat differently, and I think you're using it in a narrower sense than I mean where I probably do agree that most examples are at least a bit morally blameworthy.

000

David R. MacIver @drmaciver.bsky.social · 15/12/2024

3. I could just continue this thread into a metathreadapalooza of 100 things I could write a threadapalooza thread about...

030

David R. MacIver @drmaciver.bsky.social · 15/12/2024

1. "The legitimacy of conflict" - I keep trying and failing to articulate this in a way that I can actually explain to others, but basically the idea is "being in conflict with others is not necessarily a moral failing on either part". This claim is obvious but people keep getting it wrong.

330

David R. MacIver @drmaciver.bsky.social · 15/12/2024

I enjoyed the last threadapalooza I did, so I think I'll do another one or two in December. They're a good way of drawing out thoughts about topics, and I've got a couple of concepts that I'm trying to figure out how to articulate better. Current candidate threadapalooza topics:

160

David R. MacIver @drmaciver.bsky.social · 12/12/2024

100. And, finally, as promised: If you've got to the end of this giant rambling thread about the discord I run, and you'd still like to join it, here's an invite. Come say hi. discord.gg/AwJsHBVz4Y

discord.gg

Join the Overthinking Everything Discord Server!

Check out the Overthinking Everything community on Discord – hang out with 447 other members and enjoy free voice and text chat.

080