Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
serf
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
17 ms
·
271.
▲
by
serf
7mo ago
remote GPU compute payloads have been around a lot longer than LLMs, they're just few and far between. folding@home and other such asynchronous "get this packet of work done and get back to me' style of operations rarely care
272.
▲
by
serf
7mo ago
>Betting platforms assign highly profitable customers "concierges" who reach out and prompt them to gamble, offer incentives, and work to keep them betting. It's insidious and wrong - the platforms actively identify and ta
273.
▲
by
serf
7mo ago
that's true about assertions, but blindly saying "Fact check!" is still an attempt to offload a wished-for effort onto other people while simultaneously sowing seeds of discernment and distrust into the topic. What happens wh
274.
▲
by
serf
7mo ago
I wouldn't be so pleased with myself over such "You will get wet in a rainstorm." style predictions. truths from different angles that are at odds with one another produce mistrust and thoughts of conspiracy. We have more of
275.
▲
by
serf
7mo ago
my most altruistic view : they said it through actions. Rural areas were the last areas to join the mobile networks. This is just a practical thing though; why would you build a tower for a community of 900 people when there are still gaps
276.
▲
by
serf
7mo ago
your ethics. let's not forget that these major LLMs are all the children of corporate hyper-piracy en masse, none of them are ethical even in origin unless you're talking about the pre-product company charter kind of ethics, lik
277.
▲
by
serf
7mo ago
well moderation is ubiquitous , but yeah -- personalization/targeting/social graph are essentially the things that people expect out of a social media platform. I do personally think the karma thing is an aspect , because it
278.
▲
by
serf
7mo ago
a hammer requires an operator, so it's rarely used wrong, and if something goes wrong the operator can intervene. sometimes a thumb will be struck, but usually that will result in a painful lesson that prevents future strikes. the time
279.
▲
by
serf
7mo ago
TDD and strong goals help.. ..much like with human development.
280.
▲
by
serf
7mo ago
I have never read a snide comment on this site that i've been more repulsed by. I think because it's so specifically sharpened to stab at the software developer, my compatriot, one of the foremost primary populations here, rather
281.
▲
by
serf
7mo ago
I get your angle, but have you ever read the discourse between humans regarding fiction? I mean humans have made death threats towards other humans about whether or not Han shot first. fiction-fan-discourse is a very low bar on the rankings
282.
▲
by
serf
7mo ago
the concept of 'The average hypersonic..' makes me laugh. in actuality the concept of equating real life dollars to defense budgets makes me laugh, too. It's not really a money thing, it's a production thing; and even if
283.
▲
by
serf
7mo ago
It's a false comparison. How many MIG-25s flew over the borders of the United States mainland during the cold war? Yes the MIG-25 was a cheaper and more practical plane, but that wasn't the MO of the sr71.
284.
▲
by
serf
7mo ago
you're correct. a '12v ICE' alternator generates up to 14.8-15.2v. Most automotive stuff can operate between 9ish-16ish-v , of course totally depending on the product. of course this is just a modern interpretation. older stu
285.
▲
by
serf
7mo ago
AGI is a buzzword too, it's just differently applied. In this case it's a word that means the thing we're all developing towards apparently, but that no one actually knows how to get or even how to measure whether or not we&#
286.
▲
by
serf
7mo ago
>Rich people were always going to do what they wanted anyway, "democratizing" that doesn't make the situation better. total disagree. if you put vid gen in the hands of regular people then regular people get super-powered
287.
▲
by
serf
7mo ago
>GitHub goes down at least once a week as I said before. [0] thanks to Copilot, Tay.ai and Zoe chatbots wrecking the platform instead of humans. there are tens of thousands of stupid scripts hosted on github itself that have scheduled pr
288.
▲
by
serf
7mo ago
it's also just a numbers thing. react and more broadly JS developers are a dime-a-dozen as far as availability goes compared to winapi folks. which on one hand, good -- fuck microsoft and the monolith; on the other hand we get react st
289.
▲
by
serf
7mo ago
I like llms too, and I think they make me more productive.. but a chart of commits/contribs is such a lousy metric for productivity. It's about on par with the ridiculousness of LOC implying code quality.
290.
▲
by
serf
7mo ago
"Help me with my witch hunt!" No thanks!
291.
▲
by
serf
7mo ago
allowed shell pipes doesn't necessarily mean they want loops running them. One of the economic tuning features of an LLM is to nudge the LLM into reaching conclusions and spending the tokens you want it to spend for the question. presu
292.
▲
by
serf
7mo ago
as someone who has used codex/claude-code/opencode I can confidently say that "the very best experience on Claude" is not the one that Anthropic provides software for.
293.
▲
by
serf
7mo ago
you use profiles for that [0], or in the case of a more capable tool (like opencode) they're more confusing referred to as 'agents'[1] , which may or may not coordinate subagents.. So, in opencode you'd make a "PR M
294.
▲
by
serf
7mo ago
wasnt this sort of obvious to anyone familiar with Niantic?
295.
▲
by
serf
7mo ago
watching the difference in a non-CS versus a CS person using an LLM is all you need to do to reaffirm belief that fundamentals are still a massive benefit, if not a requirement for the deeper software.
296.
▲
by
serf
7mo ago
the holy patron saints of the internet gave us javascript and hyperlinks so that we could have both a gargantuan picture of a bee and a fast, responsive, small web transfer for discovering that gargantuan pictures' availability. peop
297.
▲
by
serf
7mo ago
I agree with the necessity of it, but I also find the argument "We do it because we do it." to be weak. We do it because we lack better methods. Cart before the horse, since those better methods are often derived from cruel rese
298.
▲
by
serf
7mo ago
the slowest DDR4 is capable of 12.6GB/s~ish per channel . nowhere near the same performance.
299.
▲
by
serf
7mo ago
>If you have 32GB Macs, and you had them 7 years ago already, you're not even remotely close to the target market for it. that market is already saturated with a zillion decent-spec chromebook style machines. The only reason the Neo
300.
▲
by
serf
7mo ago
a VM host with a windows guest and a total of 8gb of ram? Yeah you'll get the OS to run, the magic there is making either environment usable. Might be great for some web dev that needs to see what their work looks like elsewhere -- but
More ›