Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
code_biologist
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
61.
▲
by
code_biologist
10mo ago
There's plenty of overlap, but they solve different problems: flexbox when the content should control element sizing/fit, grid when the container should control element sizing/fit.
62.
▲
by
code_biologist
11mo ago
Thanks for the thoughtful comment. It's complicated. I'm a perfectionist miniature painter by hobby. I love the stories floating around about how quantity can lead to quality (photography, pottery): https://austinkleon.
63.
▲
by
code_biologist
11mo ago
Prep work sucks. In whatever field, I always love hearing from the greats how they make the hard work a little less sucky. Good tools, better workflows, figuring out what they can let slide, accepting the struggle and getting zen. The peopl
64.
▲
by
code_biologist
11mo ago
I have a print out of this on my wall. I also have pretty bad ADHD. It's been on my wall a year and I still haven't done the thing (really a series of small aversive, timeliness and administration heavy things). I oscillate betwee
65.
▲
by
code_biologist
11mo ago
Love it. This has Cobra Effect style perverse incentive written all over it. You'd be shocked how quickly you can get a big bag of blood vials if you know the right people.
66.
▲
by
code_biologist
11mo ago
100% — editing is trivial to argue against. Embryonic selection is much more difficult to mount clear, decisive arguments against. Let's make sure we don't end up with net skewed sex selection, and not screw up gene/phenotype
67.
▲
by
code_biologist
1y ago
Reading Erowid trip reports persuaded me to try high dose tripping on Robitussin (DXM). From Wikipedia: The first plateau (1.5 to 2.5 mg per kg body weight) is described as having euphoria, auditory changes, mild stimulation, and change in
68.
▲
by
code_biologist
1y ago
I've seen a lot of young people (teens especially) cross active streets or cross in front of Waymos on scooters knowing that they'll stop. I try not to do anything too egregious, but I myself have begun using Waymo's conserva
69.
▲
by
code_biologist
1y ago
I'm reminded of this Steve Jobs story: So it's the MacBook Air guy's turn. He comes in and places his prototype down in front of Steve. Steve opens the lid. Two seconds later he picks up the laptop and heaves it so hard it s
70.
▲
by
code_biologist
1y ago
100%. Not eating spoonfuls of Crisco/shortening like peanut butter is an easy life change. "Reduce chronic stress" is non-trivial.
71.
▲
by
code_biologist
1y ago
I would second your suggestion to investigate bupropion. More than a decade ago I was suffering from severe depression. Psychiatrists with meds were an easy-to-access front line treatment, but I was very concerned about SSRI side effects, p
72.
▲
by
code_biologist
1y ago
Great post. I too have a G305 with Kailhs. I felt nuts replacing the switches, and I'd never desoldered anything in my life before, but gosh the silent is so good.
73.
▲
by
code_biologist
1y ago
I recently got a newer iPhone and moved to iOS 18 with the hardware change. I had to watch some youtube tutorials explaining navigation and swipe locations. Over and over I've had the thought "this never would have flown under Ste
74.
▲
by
code_biologist
1y ago
Eh, as long as it isn't life or death, I think allowing direct consumption and explicitly agreeing that breakage is a consumer problem is better for most business use cases (less code, easier to maintain and evolve). If you make a brea
75.
▲
by
code_biologist
1y ago
As a person who easily falls for "preparing" over doing the work, I committed to the hipster PDA [1] as an antidote to this for many years: "the Hipster PDA comprises a sheaf of index cards held together with a binder clip.&q
76.
▲
by
code_biologist
1y ago
Right, so latent deceptiveness has to be favored in pretraining / RL. To that end, it needs to be: a) useful to be deceptive to achieve CoT reasoning progress as benchmarked in training b) obvious deceptiveness should be "selected
77.
▲
by
code_biologist
1y ago
They'll improve, but current LLMs (o3, Gemini 2.5 Pro, lesser ones) are also terrible at suggesting non-obvious alternative solutions on their own. You can sometimes squeeze them, adding each obvious but poor approach to a list of &quo
78.
▲
by
code_biologist
1y ago
Can I ask why you use it? Honestly just for fun? I think girl stuff is cool and my partner indulges me. It does truly help with emotional/sexual/physical planning, but a very biological embodied flavor of fun is my real reason.
79.
▲
by
code_biologist
1y ago
My use case is mostly creative writing. IMO 2.5 Pro 03-25 was insanely good. I suspect it was also very expensive to run. The 05-06 release was a huge regression in quality, most people saying it was a better coder and a worse writer. They
80.
▲
by
code_biologist
1y ago
I would like the option to pay for the unquantized version. For creative or story writing (D&D campaign materials and such) quantization seems to end up in much weaker word selection and phrasing. There are small semantic missteps that
81.
▲
by
code_biologist
1y ago
I experience the same but I've always been an extremely strong speller. I think it's a biased viewpoint. I remember the kids in grade school who really struggled, and I've always wondered how that same group fares these days
82.
▲
by
code_biologist
1y ago
It's been interesting to watch this play out in microcosm in different spaces. Danbooru and Gelbooru are two anime image boards that banned AI image content, largely to their benefit in my opinion. Rule34 is a similar image board that
83.
▲
by
code_biologist
1y ago
I don't see how being forced to upgrade all consumers is a good thing. It forces implementers of broad or disruptive API or technical changes to be responsible for the full consequences of those decisions, rather than the consumers o
84.
▲
by
code_biologist
1y ago
I think your parent comment is citing that as an example of why livebench is no longer a good benchmark. That said, the new Flash is very good for what it is, and IMO after the Pro 05-06 nerfs the two models are much closer in performance f
85.
▲
by
code_biologist
1y ago
That's the neat part. They don't. Either the broken consumer updates their use, you update it for them to get your change shipped, or you add some backwards compatibility approach so your breaking changes aren't breaking.
86.
▲
by
code_biologist
1y ago
fuzzycomplete - https://github.com/Query-farm/fuzzycomplete "This fuzzycomplete extension serves as an alternative to DuckDB's autocomplete extension, with several key differences: ..." lindel - https:&
87.
▲
by
code_biologist
1y ago
This Ars Technica post is pretty good: https://arstechnica.com/ai/2025/05/why-google-geminis-pokemo... Sad to see it didn't get HN traction.
88.
▲
by
code_biologist
1y ago
Using an LLM to do schoolwork is like taking a forklift to a gym where you're told the goal is to be healthy and strong, but they can't really stop you from using a forklift, and jobs and compensation are given out according to ho
89.
▲
by
code_biologist
1y ago
Just an app dev, not a geospatial expert, but reprojection always seemed like something a library should handle under the hood unless one has specific needs. I'm used to the ergonomics / moron-proofing of something like Postgis&#x
90.
▲
by
code_biologist
1y ago
At my last startup, I think our board and observers liked hanging out more than they liked talking about the company. It wasn't perfect, but it's their money so I wasn't going to complain.
More ›