bnjmn

8 ms·

1.
by bnjmn 5mo ago
"Context and intent cannot be known" seems like a bit of an overstatement? A qualifying clause like "in all cases, with complete confidence" would allow for the possibility of alignment in some cases (yay), but not alway
2.
by bnjmn 6mo ago
This game was the first software I ever modified! I added an easter egg where if you spent all your money immediately on cocaine, its price would reliably skyrocket, so you could go ahead and pay back the loan sharks, or buy lots more drugs
3.
by bnjmn 11mo ago
> Humanoid robots are handicapped by their similarity to humans which is really an artificial constraint. YES, and I wish people would stop pretending we've unlocked some new generality by promoting generic humanoid robots over task
4.
by bnjmn 11mo ago
Here's a use case that seems more science fictional to me (as the parent of a 2yo) than warp drive: a robot that can gently restrain an uncooperative human baby while changing its diaper, with everything that entails: identifying and e
5.
by bnjmn 1y ago
Among other general advice, my CLAUDE.md insists Claude prove to me each unit of change works as expected, and I'm usually just hoping for it to write tests and convince me they're actually running and passing. A proof assistant s
6.
by bnjmn 1y ago
They say the e-ink display has "Unmatched Speed, Never Seen on ePaper" so it would nice to know the actual refresh rate. This is not an endorsement, but https://daylightcomputer.com/ claims 60fps, so that's t
7.
by bnjmn 2y ago
Thanks for these detailed replies! Now I really want to read your paper.
8.
by bnjmn 2y ago
I fully agree with your final statement, but needing to constrain the problem in an artificial way to prove it's NP-complete doesn't mean the constraint was justified or realistic, because then you've only proved the constrai
9.
by bnjmn 2y ago
> We still do not know, for instance, what makes a good tokeniser (Gowda and May, 2020; Cognetta et al., 2024): which characteristics should its produced subwords `s` have to be a good starting point for language modelling? If we knew th
10.
by bnjmn 3y ago
On any macOS computer (or replace /usr/share/dict/words with your own word list): grep '^[qwertyuiop]*$' /usr/share/dict/words | \ awk '{ print length(), $0 }' | \ sort -n
11.
by bnjmn 3y ago
Are there any new cars / car brands that credibly promise not to track their drivers? Any car with a network connection for software updates seems likely to be harvesting driver data, or is at least capable of doing so.
12.
by bnjmn 6y ago
Mostly because they admit those results are not statistically significant (for example, at the bottom of the summary diagram). That said, I admire their honesty, and I hope these results suggest other kinds of experiments to other researche
13.
by bnjmn 6y ago
Although these results are not exactly impressive or compelling (they don't make me want to change my exercise habits), it's reassuring to see researchers go through with publishing underwhelming results, rather than cherry-pickin
14.
by bnjmn 6y ago
"Potentially" is an understatement! A much better take, IMO.
15.
by bnjmn 6y ago
This isn't exactly what you asked for, but it seems pretty relevant to your stated interests: https://www.jasondavies.com/maps/voronoi/capitals/
16.
by bnjmn 7y ago
I don't think you can have a proper Merkle tree if cycles are possible, and cycles definitely are possible in this AST representation: https://www.unisonweb.org/docs/faq#how-does-hashing-work-for...
17.
by bnjmn 7y ago
My thoughts as well!
18.
by bnjmn 7y ago
you're right, I should have replaced "intuitions" with "observations" instead of waffling
19.
by bnjmn 7y ago
Scientific understanding (defined by the ability to make accurate predictions) can be (must be?) anchored in both the math and physical intuitions/observations, and it's very difficult to bridge the two without resorting to natu
20.
by bnjmn 7y ago
One of my favorite takes on this topic: "From... elementary theories we build up descriptions of more and more complex systems. But in all these efforts we take for granted that we may use any language we wish and as many [languages] a
21.
by bnjmn 7y ago
Thanks!
22.
by bnjmn 7y ago
Speaking as a maintainer of large open-source projects, I know what it's like to get stuck working on a release branch (which is sometimes just the `master` branch) that takes forever to become an official release. I'm not sure if
23.
by bnjmn 7y ago
Here's a comment from Sebastian M (React core team member) about why they chose not to use generator functions: https://github.com/facebook/react/issues/7942#issuecomment-2...
24.
by bnjmn 7y ago
Yes! I was also curious about this, because I'm sure you could make generators/yield work for asynchronous interruptible rendering, but here's what the React team has to say: https://github.com/facebook/r
25.
by bnjmn 7y ago
Here's their official position on generators: https://github.com/facebook/react/issues/7942#issuecomment-2...
26.
by bnjmn 7y ago
Does anyone know why they chose Node.js 12.8.1 instead of 12.13.0, the first LTS release on the 12.x line?
27.
by bnjmn 7y ago
A potentially less labor-intensive way to estimate animal population is to capture a sample, tag them, release them back into the wild, and then later take a second sample and see how many of those animals are tagged: https://en.
28.
by bnjmn 7y ago
Thanks, this is a good take. I would add: now it's everyone else's job to make those smaller, more focused libraries work together. The time wasted is unfathomable, but the blame for misconfiguration now falls on individual develo
29.
by bnjmn 7y ago
> Does this mean Meteor is dead? Absolutely not! We depend on it internally (engine.apollographql.com is a Meteor app), and Galaxy hosting revenue is THE reason we didn't have to raise money until we felt like it, which is why we go
30.
by bnjmn 8y ago
Looking through the supported APIs ( https://docs.aws.amazon.com/documentdb/latest/developerguide... ), it appears DocumentDB has no support for Mongo's oplog ( https://docs.mongodb.com/manual&#x

More ›