Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
stymaar
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
14 ms
·
271.
▲
by
stymaar
3mo ago
Do you have any aircraft on mind that can unfold its wing while flying? (I'm aware of variable-sweep wing which can reconfigure in flight, or foldable wings that will be fold on the ground before parking).
272.
▲
by
stymaar
3mo ago
This! And this is why the “crypto” world is full of scams and heists, in a way that anyone with a brain could have foreseen from the start.
273.
▲
by
stymaar
3mo ago
Until customers don't want to but because they are afraid they have no way to get their money back if anything goes wrong. (Chargeback exists for a reason, and the reason isn't to piss off businesses).
274.
▲
by
stymaar
3mo ago
This was thr original intent of thr Satoshi paper, except BTC is probably the worst payment network you can think about, which is why it failed to gain any use as such over the past 18 years…
275.
▲
by
stymaar
3mo ago
> The defining characteristic of a commodity is that it is fungible: a gallon of oil is a gallon of oil; a ton of copper is a ton of copper; a bushel of wheat is a bushel of wheat. The concept of “commodity” as defined above is a model,
276.
▲
by
stymaar
3mo ago
There has always been a magic money tree, but it's not for everyone.
277.
▲
by
stymaar
3mo ago
This the proof that guillotine is useful, if not to be used, to make the elite behave in way that avoid suffering it.
278.
▲
by
stymaar
3mo ago
Rust as a language is irrelevant to this discussion, the problem is that bun was vibe coded by a single dude without any open source community involvement. “bun the open source project” is basically dead at that point: don't expect any
279.
▲
by
stymaar
3mo ago
Its not a typo, someone coined that during the DeepSeek R1 hype period and I kept using it since then. I totally agree with you on the fact that it's not morally any different than pre-training. IMHO we should have a legislation that f
280.
▲
by
stymaar
3mo ago
> Kimi K3 reproducibly identifies itself as Claude It could also be have been trained from collected response datasets. Claude got caught several time responding it was ChatGPT or even Deepseek and I don't think Anthropic has been d
281.
▲
by
stymaar
3mo ago
> Germany manages free speech and democracy as well as it does producing software. That sentence works suprising well for the US too: enshitified and completely controlled by a cartel of trillion dollar corporations.
282.
▲
by
stymaar
3mo ago
True. I don't understand why they overwrite the date at all actually, for either the post or the comments, that sounds like the hackiest way of implementing the “second chance pool” feature.
283.
▲
by
stymaar
3mo ago
Second chance pool, most likely.
284.
▲
by
stymaar
3mo ago
What happens if you get margin called in this scenario?
285.
▲
by
stymaar
3mo ago
But is it load-bearing?
286.
▲
by
stymaar
3mo ago
> Directly not to care because they lost in court. I'm pretty sure the illegal sport streaming websites didn't stop doing that just because it became illegal, otherwise they could have stopped their activity altogether while th
287.
▲
by
stymaar
3mo ago
I don't think it's a very good remark, as there's significantly less email spam than 20 years ago. Another example is ad-blocker-blocker. There was a little bit of an arm race between ad blockers and advertisers in the middle
288.
▲
by
stymaar
3mo ago
> Over-using em-dash or whatever isn't the thing that maximizes engagement. It's the thing that minimizes the loss during the RLHF phase, and the RLHF phase is the one that's aimed at maximizing engagement (it's liter
289.
▲
by
stymaar
3mo ago
> The first is that it would still misclassify human-authored text written under the same incentive, and most people have various incentives to "maximize engagement". The thing is, humans are significantly worse at maximizing n
290.
▲
by
stymaar
3mo ago
> but it's a bad fiction to perpetuate that any of this is anything more than tarot card reading. Hard disagree. LLMs (especially base ones, that only received pre-training) can produce output that is undistinguishable from human wr
291.
▲
by
stymaar
3mo ago
Irrespective to the technical merits of both language, moving from a stable language to a pre-1.0 one that just lost his most popular open source project is a wild move.
292.
▲
by
stymaar
3mo ago
> AI prose is better than the average human There's no objective measurement of “better”, which in itself refutes your argument, but I personally feels AI prose to be very annoying for mostly two reasons: - it's the same everyw
293.
▲
by
stymaar
3mo ago
> True capitalism has never been tried “True communism has never been tried”
294.
▲
by
stymaar
3mo ago
> I’m still having to deal with making some sort of legal agreement with EVERY website I visit over cookies constantly… I just don’t care anymore / click whatever. That's the problem when you want to be pro-business while keepi
295.
▲
by
stymaar
3mo ago
We've built an entire civilization (the industrial civilization) around the idea that competition is where project comes from, I don't think we have the necessary mental framework, let alone the institutions, to make that kind o
296.
▲
by
stymaar
3mo ago
> American Marxist degrowth mindsets. A literal Communist dictatorship is winning the race thanks to their 5-years plans yet you still think the problem is “too much Marxism in the US”. What an utterly deranged take.
297.
▲
by
stymaar
3mo ago
It's not even the worse part. Journalists have used the "it is not A, it is B" for a while and it never was as bad as AI doing it. The key problem is that AI have no “idea” of how to build an argument, so it just rephrases th
298.
▲
by
stymaar
3mo ago
> Check out 2015, it had way hotter temperatures in November That's what El Niño is about: instead of declining after a peak in June, temperature plateaus and then rise again later in the year. So the fact that the Pacific is much h
299.
▲
by
stymaar
3mo ago
Maybe, but it's refreshingly different from the industrialized junk prose we get to read more and more.
300.
▲
by
stymaar
3mo ago
> No. The (standard) deviation is 3.5. 3.5 what, according to you? You're reading this graph wrong: we're currently 3.63 standard deviation above the mean. It's clearer on the original article[1] that this AI-generated blo
More ›