unholiness

6 ms·

31.
by unholiness 4mo ago
Yeah, this made it basically clickbait for me, in terms of time I wasted with the wrong expectation. The lack of downvotes on posts on HN has always felt like more of a bug than a feature to me.
32.
by unholiness 4mo ago
Subagents are a helluva drug.
33.
by unholiness 4mo ago
As the 36 millionth person building something similar, I wish[0] there were better info out there on what works well. I can understand why, but it's still frustrating seeing on the one hand how deeply helpful the flywheel of this type
34.
by unholiness 4mo ago
The Trustees of Reservations, in the Boston area, are a great example of this working well.
35.
by unholiness 4mo ago
If it were, I can in theory see situations where improving content cleanliness is worth blowing away the KV cache. But I absolutely can't see how feeding the entire context into a more expensive model multiple times per task, just to p
36.
by unholiness 5mo ago
Gene drives are such amazing and such frightening technology. No one puts them in the same conversation as nukes or engineered pandemics, but they share the same pattern of "technology improvements giving smaller and smaller actors glo
37.
by unholiness 5mo ago
I'm having a hard time squaring this view with the prediction markets. 500k+ of volume on Polymarket[0] had Elon losing at only even odds until March, 2:1 until May 15th, and 3:1 before the verdict dropped. It seems to have been widely
38.
by unholiness 5mo ago
I just open photos.google.com and grab them. No need to fiddle on my phone. When on wifi, the photo backup upload starts immediately. If it doesn't (possibly due to your settings, this used to be my issue) you can manually open the pho
39.
by unholiness 5mo ago
https://xkcd.com/277/
40.
by unholiness 5mo ago
I think the answer is practice, for a few reasons. One is obvious: conversation is a skill. Just like a novice chess player can spend 5 seconds figuring out which squares the knight can move to while an intermediate player spots a fork to f
41.
by unholiness 6mo ago
> That's getting solved already in china leading labs, and bottlenecked by the lack of good training data, which china is solving with mass labor. What work are you referring to here?
42.
by unholiness 6mo ago
More accurately, binary search is optimal only if you cannot determine distance between the data points (you can compute `<` but not `-`). It's inaccurate to say it's optimal if you know "nothing" about the distribu
43.
by unholiness 6mo ago
> Here is the legal baseline, in plain terms: This particular AI-ism really encapsulates what annoys me about some AI-isms. I don't mind the delves and the em-dashes that just give away the AI source of what otherwise might be good
44.
by unholiness 6mo ago
It applies more to design software, where a user is creating durable things and needs to understand those things themselves. Google Maps is more of an agent: It's responsible for understanding its own complexity and answering your quer
45.
by unholiness 6mo ago
I cringed very hard in the slerp example seeing `acos(dot(a,b))`. Clamping to [-1,1] to avoid NaNs still gives you bad answers and numerical sensitivity around small angles. acos and asin in general lose ~half your sig figs around their sin
46.
by unholiness 6mo ago
I love the fractal nature of this, where the big shape of one two three four... is then roughly repeated both on a slower scale (twenty thirty forty...) and on a narrower scale (twenty-one twenty-two twenty-three twenty-four...). I'm n
47.
by unholiness 7mo ago
I discovered them last year on Substack and they quickly became a priority read. A sort of Quanta for biology, taking time to explain enough for a popular audience but keeping technical rigor deep into some fascinating topics. Some highligh
48.
by unholiness 7mo ago
Solidworks has VBA macros, which (on top of being poorly documented and unstable) subvert to the whole benefit of parametric CAD. Once you're creating features with a macro, you naturally want to edit them, but you also naturally want
49.
by unholiness 7mo ago
The relevance is that Claude made this cheat sheet.
50.
by unholiness 7mo ago
This is true if the milk is in the fridge the whole time. With the milk out the whole time, it's nearly the same, exact answer depends on the geometry of both containers.
51.
by unholiness 7mo ago
He's definitely conflating two things: watches changed from utilities to fashion items, and watchmakers changed from engineering to branding. In this example, the one change caused the other, but it's just one example. It's n
52.
by unholiness 8mo ago
I like it! One big annoyance with the power-ups is that the failure condition is checked before you can use them. It's particularly painful since they all replace the current piece, so they seem tailor-made to get rid of a nasty piece
53.
by unholiness 8mo ago
No, their Responsible Scaling Policy and their government contract are not related. The RSP governs how Anthropic itself behaves w/r/t developing, testing, and releasing new models. The contract was signed with stipulations around
54.
by unholiness 8mo ago
One difference is the very real possibility that AI will not just be a "tool for humanity", but a collection of actors with real power and goals. Robert Miles has an approachable explanation here: https://www.youtube.co
55.
by unholiness 8mo ago
Not sure this is necessarily for faster growth. Riding out the AI bubble's rise and/or its bursting will each present a lot of need for capital and a lot of barriers to raising it. They're not an AI company but they obviously
56.
by unholiness 8mo ago
Security is a classic example of a public good where this doesn't work well. The cheapest ways to secure an airport (sharing queues, staff, protocols, machines, training, threat models) are going to also benefit those who opt out, crea
57.
by unholiness 10mo ago
Loving the blog in both style and content, hope to have time to read more in the future! A random note in case Non-Zero-Sum James is looking: It's frustrating that reading footnotes[0] requires scrolling back and finding your previous
58.
by unholiness 2y ago
The way I see it: The reasonable things that continue happening each day in our universe would be extremely unlikely if we are just Boltzman brains. Every bit of sensible reality would be coincidental. The very continuance of that reality
59.
by unholiness 2y ago
As excited as I am about this jump from the fuzzy NeRFs/gaussian splatting to real meshes, I'm not holding my breath for BREP generation. Mesh to BREP has always been fraught because for anything beyond "find the cylinders&qu
60.
by unholiness 2y ago
That's right. While a day passes on earth, a day minus 58.721 μs passes on the moon (which is moving faster than the Earth). This multiplies, so after a million days a clock on the moon will read 58 seconds behind a clock on the earth.

More ›