Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
skissane
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
60 ms
·
871.
▲
by
skissane
2y ago
> Again, your claim here is that it's typical and predictable that new administrations conduct ideological purges on the civil service. No, I'm not denying this is reaching a level which hasn't been seen before. But, perce
872.
▲
by
skissane
2y ago
More likely we’d all be using FreeBSD
873.
▲
by
skissane
2y ago
Well, humans already produce "an innocuous, test-passing function that's catastrophically wrong" even without LLMs involved. So, the real question is, will LLM adoption result in a significant increase in such incidents? I do
874.
▲
by
skissane
2y ago
> and chucking out a significant chunk of work they've done It is unclear how much of the actual work they've done is being "chucked out". 18F did work for various federal agencies, and whatever code 18F wrote for its
875.
▲
by
skissane
2y ago
> And now neither of them exist (the vast majority of what once was USDS is gone, and what remains has been converted into "DOGE"). Do any of the old USDS staff survive? I don't know. USDS acting administrator, Amy Gleason
876.
▲
by
skissane
2y ago
Not strictly speaking a government agency, but what about the Judicial Procedures Reform Bill of 1937? I think there was another possible reason for getting rid of 18F, separate from any concerns about its political culture – there was a lo
877.
▲
by
skissane
2y ago
Digressing a bit: is there any correlation between Bangaldeshi-Americans preferences in US politics (Democrats vs GOP) and their preferences in Bangladeshi politics (Awami League vs BNP vs Yunus)? Or are those largely orthogonal dimensions?
878.
▲
by
skissane
2y ago
If a government agency has a culture which appears to lean strongly in one political direction, it is unsurprising that when the opposite political persuasion gets into power, the agency becomes a target. Traditionally how civil servants ha
879.
▲
by
skissane
2y ago
The tweet the GP linked – https://x.com/lukerosiak/status/1885523747425399247 – mentioned they wrote a Slack bot to lecture people about using inclusive language. I don't think that is average for US tech wor
880.
▲
by
skissane
2y ago
I think you are right that the US constitution, as originally written, doesn't provide any internal "checks and balances" on the executive branch, other than the President. Congress and the judiciary act as external checks an
881.
▲
by
skissane
2y ago
Can someone confirm if this is an accurate summary of the article: "CLOS dependent maintenance protocol" allows you to run code when a class definition (or one of its ancestors or descendants) changes?
882.
▲
by
skissane
2y ago
They had a rather progressive-coded internal culture, [0] which I think made them very likely to have been attacked by a Republican administration, even if they'd never worked on direct filing at all. Contrary to the media narrative wh
883.
▲
by
skissane
2y ago
Read it in context: Musk's tweet saying 18F "has been deleted": https://x.com/elonmusk/status/1886498750052327520 Was replying to ALX's tweet calling 18F "far left": https:/
884.
▲
by
skissane
2y ago
I think it is unlikely (of course not impossible) an LLM would fail in that way. The underhanded C contest is not a case of people accidentally producing highly misleading code, it is a case of very smart people going to a great amount of e
885.
▲
by
skissane
2y ago
The fact they make mistakes like this implies they probably don’t, since surely steps like that would catch many of these
886.
▲
by
skissane
2y ago
> But the Google Search LLM provides flagrant, laughable, and even dangerous misinformation constantly. It’s a public service: helping the average person learn that AI can’t be trusted to get its facts right
887.
▲
by
skissane
2y ago
I think a lot of these issues could be avoided if, instead of just a raw model, you have an AI agent which is able to test its own answers against the actual software… it doesn’t matter as much if the model hallucinates if testing weeds out
888.
▲
by
skissane
2y ago
> China isn't sending a "small number"; casualties alone have supposedly reached into the hundreds I think the word "supposedly" is important here – I don't think we have any hard data on how many Chinese vo
889.
▲
by
skissane
2y ago
> (1) Chinese nationals have served, indicating that Beijing at least tacitly approves mercenary actions I wouldn't assume a small number of Chinese nationals volunteering to fight for Russia means China approves of their actions. S
890.
▲
by
skissane
2y ago
You can open a US dollar account in a non-US bank. Some banks don't offer this service to ordinary private citizens (others do), but I think pretty much all will if you are a (non-sanctioned) business or foreign government or high net
891.
▲
by
skissane
2y ago
Schwitzgebel‘s isn’t assuming any property of consciousness is transitive. Rather what he is saying is this: given many candidate substrate-independent materialist definitions of consciousness, if the property is true of individual humans a
892.
▲
by
skissane
2y ago
> The argument makes assumptions that consciousness is transitive or "sticky" in some sense, where if a property applies to a part of the system it must also apply to some aggregate of those parts. I think the argument is reall
893.
▲
by
skissane
2y ago
I think this article has the problem that it is addressing an interdisciplinary topic with too much focus on only a single discipline, which can be a sign of lacking sufficient diversity in disciplinary background of peer reviewers. Scott A
894.
▲
by
skissane
2y ago
It has never paid my bills, in that I've never worked for an analyst firm. My bills have been paid by working for vendors, where I have seen how sales and marketing use their reports in action. I have seen the amount of effort engineer
895.
▲
by
skissane
2y ago
> Better let Nature know their Peer Review committee screwed up then! The article you cite [0] is labelled as "opinion". The standards for peer review of opinion articles in scientific journals are a lot lower than those for or
896.
▲
by
skissane
2y ago
A lot of the people paying for these analyst firm reports are sales people-so they can pass them on to their customers/prospects (to legally do that you often have to pay extra for “redistribution rights”)… and then the customer/p
897.
▲
by
skissane
2y ago
> These clowns want $2500 goddamned american dollars for the privilege of reading their bloviations on this topic, which i absolutely will not pay. You aren’t the target market for their “bloviations” - they are targeted at executives, a
898.
▲
by
skissane
2y ago
> but greater is a geographic term and doesn't mean "better" instead it's closer to "wider" or "covers more area." You are confusing a specific geographical sense of “greater” (e.g. “greater New Yo
899.
▲
by
skissane
2y ago
There's a great philosophy paper making this argument: Schwitzgebel, Eric. “If Materialism Is True, the United States Is Probably Conscious.” Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, vol
900.
▲
by
skissane
2y ago
There is such a thing as a philosophical mistake – almost everyone agrees that logical positivism was a failure, despite the fact that there was a period in the 1950s when it was all the rage in philosophy departments in the English-speakin
More ›