Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
groby_b
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
14 ms
·
271.
▲
by
groby_b
1y ago
Curious: Do you also laud your compiler for particularly good optimizations?
272.
▲
by
groby_b
1y ago
It's not that the terms are unpopular, it's that every system that doesn't have strong capitalist roots has lost out to more capitalist systems. Nothing wrong with alternatives, but for that we need to let go of "let
273.
▲
by
groby_b
1y ago
Thank you, appreciated - and the commentary as well.
274.
▲
by
groby_b
1y ago
> You can probably give away free school lunches for an entire generation of children with that $10 billion in Louisiana The annual cost of the National School Lunch Program is $18B, so, no.
275.
▲
by
groby_b
1y ago
Good journalism is stating the relevant facts concisely, not blabbering for 30 minutes.
276.
▲
by
groby_b
1y ago
If I were in Dreamwidth's shoes, I'd be very much concerned with minimizing legal exposure, not number of users excluded. At 10k/user*day, it's a reasonable choice to block as broadly as makes sense. Tough for the neighb
277.
▲
by
groby_b
1y ago
Or you could give the key points instead of asking people to consume 2 hours of content.
278.
▲
by
groby_b
1y ago
I feel like any article on burner phones that glosses over acquisition with "buy phone and service in cash" misses the point. Buying a phone anonymously is much harder than "just cash". Most places demand name & ad
279.
▲
by
groby_b
1y ago
One, not everybody gets 500K TC. Two, several tens of thousands are in the 5%-10% range. Hardly "peanuts". But I suppose you'll be happy to hear "no raise for you, that's just peanuts compared to your TC", righ
280.
▲
by
groby_b
1y ago
"even the rich FANG types" I think you wanted to say "especially". You're exchanging clearly measurable amounts of money for something extremely nebulous like "developer productivity". As long as the perso
281.
▲
by
groby_b
1y ago
I did not know this indeed. But, of course, that means signing up and spreading your medical info even further. (Plus, in some cases, more money) For folks who are also looking, the search term "adhd meds call around service" seem
282.
▲
by
groby_b
1y ago
This would make sense if it were a single issue. (And then, limited sense - even quite red places like Texas invest in solar because they need it. Coal is a regional issue, not a national one) If you see it in light of all the other actions
283.
▲
by
groby_b
1y ago
Honestly, as somebody who filled out way too many CBT forms - there's a good chance you only get the score improvement because people don't want to get the provider in trouble with their employer. The scoring sheets are, pretty mu
284.
▲
by
groby_b
1y ago
Your cynical hunch is very much wrong. That change was easier to swallow on massive capitalization than for smaller businesses.
285.
▲
by
groby_b
1y ago
This doesn't even account for the perma-drama due to the artificial shortage generated by the DEA. Getting your prescription filled, if you manage to wrangle one, often requires hours of calling around which pharmacy can fill said pres
286.
▲
by
groby_b
1y ago
WFH also made it massively easier for people to slack off/pretend to do multiple jobs. And life was disproportionately more difficult for more junior folks. WFH was far from an unalloyed good that only got abandoned "because it ma
287.
▲
by
groby_b
1y ago
Except.... "You produce timely results" and "your colleagues are aware what you do" is something that does matte, a lot. And while PR counts and friends aren't a metric to measure people by, that sure are a signal t
288.
▲
by
groby_b
1y ago
It's not "tired" to see if something is actually relevant in context. LLMs do not exist as marvel-qua-se, their purpose is to offload human cognitive tasks. As such, it's important if something is a commonly shared fai
289.
▲
by
groby_b
1y ago
I think you're wrong with those groups. I'll use myself as counterexample, but I know a sufficiently large number of people like me who also love AI to suggest the pattern's wrong. Programming for 4 decades, happy to language
290.
▲
by
groby_b
1y ago
> read before u run Lovely sentiment, not applicable when you actually work on something. You read your compiler/linker, your OS, and all libraries you use? Your windowing system? Your web browser? The myriad utilities you need to g
291.
▲
by
groby_b
1y ago
That's not necessarily what matters. What matters is if there is a shared representation space across languages. If there is, you can then (theoretically, there might be a PhDs and a Nobel or two to be had :) separate underlying struct
292.
▲
by
groby_b
1y ago
> how do you know how much is correct Because it's a budget. Verifying them is _much_ cheaper than finding all the entries in a giant PDF in the first place. > the butterfly effect of dependence on an undependable stochastic syst
293.
▲
by
groby_b
1y ago
If you are professionally responsible for infrastructure, and you haven't thought very hard about the "how do we rebuild from scratch" case, you're committing professional malpractice. The problem is environments like ho
294.
▲
by
groby_b
1y ago
Hey, I'm not arguing against having kids. I think they're awesome, too, even though life didn't allow them for me. I'd just like discussions to be rooted in actual data instead of sentiment.
295.
▲
by
groby_b
1y ago
The US sees about 20K teen pregnancy abortions. That's probably not why the number of births is way down. Number of births in the US are ~3.6M right now. We also have 1M abortions per year. That's - if abortions were the sole prob
296.
▲
by
groby_b
1y ago
About 100%? When do you write code that doesn't need to search? (Unless you hand it all to the database, in which case... sure, you're good)
297.
▲
by
groby_b
1y ago
Come on, write couleur and let them eat cake ;)
298.
▲
by
groby_b
1y ago
That was not the argument. The argument was that their case "lost six-figure job, working as waiter now" was covered by GDP/capita, and it is. Applebee's doesn't make the same revenue per capita as a place that hand
299.
▲
by
groby_b
1y ago
It still misses the fact AI is nowhere close to self-improvement. In fact, there was a paper out on Friday that shows they're impressively bad at it: https://arxiv.org/abs/2506.22419
300.
▲
by
groby_b
1y ago
At least for me, the core criticism of AI 2027 was always that it was an extremely simplistic "number go up, therefore AGI", with some nice fiction-y words around it. The scribble model kind-of hints at what a better forecast woul
More ›