Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
dTal
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
51 ms
·
781.
▲
by
dTal
3y ago
This is one of those times where the math is so simple that it's baffling that anyone misses it. When a private company must 1) do the same job as a public service while also 2) coming up with surplus bags of cash to hand over to inves
782.
▲
by
dTal
3y ago
You have so many more practical options for even amateur security auditing on Linux. You can trivially spin up a temporary OS/chroot/container and run your specimen inside it, or run it under strace and log every system call, or
783.
▲
by
dTal
3y ago
It's not so much that systemd is designed to lure in casual users, it's that it (like other software on the list) reflects an attitude that separates "developers" who decide how a system should work, from "users&quo
784.
▲
by
dTal
3y ago
Systemd definitely belongs in the list, even if some admins like it. "Windowsification" is a admittedly vague concept, but tends to cover software that is gratuitously different, not a team player in the bazaar, prone to making la
785.
▲
by
dTal
3y ago
Who are you, or anyone else, to decide what is and isn't "culturally relevant"?
786.
▲
by
dTal
3y ago
Your faith in humanity is admirable but sadly misplaced.
787.
▲
by
dTal
3y ago
Maybe this was once true, but nowadays I'm afraid the etiology is rather more complicated than that - flat eartherism is now full-on cult territory. I recommend the documentary "Behind the Curve".
788.
▲
by
dTal
3y ago
I see no evidence of woo-orientation slowing down - Deepak Chopra has seemingly never been more popular - and angry backlash is a perfectly understandable response to swathes of people being taken in by lies with spiritual overtones. We sho
789.
▲
by
dTal
3y ago
That's victim blaming. They didn't overtly complain about an issue with the project, they were asking for usage guidance. This is the kind of contextual inference which language models are supposed to be (relatively) good at. The
790.
▲
by
dTal
3y ago
It's suggestive that a potentially embarrassing chatlog of GPT-4 hosted on an OpenAI domain got taken down within the hour of it going viral, 3 weeks after being posted.
791.
▲
by
dTal
3y ago
And here it is in Unicode glory: ∠τ = 1
792.
▲
by
dTal
3y ago
e^pi + 1 = 24.14069263...
793.
▲
by
dTal
3y ago
That we should consider all of the animals we have killed and enslaved over the millennia as people .
794.
▲
by
dTal
3y ago
This is also why it's a bad idea to read propaganda outlets to "stay more informed".
795.
▲
by
dTal
3y ago
> I also carry the common scepticism caused by philosophers not getting literally anything right about the 20th-century-physics picture of the world right ahead of time Well, Democritus did. But nobody paid much attention. Which is the f
796.
▲
by
dTal
3y ago
I fail to see the analogy between "aliens must exist" (a statement for which there is no evidence) and "the submersible imploded" (which is substantiated by debris). The syllogism is simple: * submarines make loud noises
797.
▲
by
dTal
3y ago
In addition, note that despite losing contact while the sub was on its way down, they did not report it missing until it was overdue after the full mission length. That indicates a certain blase, routine attitude to communication loss.
798.
▲
by
dTal
3y ago
The following things are all true: * the term "open source" predates the term "free software" in the FSF sense * "open source" was proposed as a synonym for "free software" in the late 90s * liguistic
799.
▲
by
dTal
3y ago
People always mention Frink and GNU Units, but the workhorse of desktop unit-aware calculation is, in my opinion, Qalculate. In addition to being a capable scientific calculator in its own right, its killer feature is the ability to throw &
800.
▲
by
dTal
3y ago
Antidepressants also sometimes give people the get-up-and-go to commit suicide. But in general it's good to have motivation to do things.
801.
▲
by
dTal
3y ago
Theft of a physical object must also be intentional. Wage theft is very much alike to theft, although it would be more properly termed "labor theft".
802.
▲
by
dTal
3y ago
Why bother with the micromanagement? "Computer, waste time commenting on Hacker News for three hours."
803.
▲
by
dTal
3y ago
Maybe it's for if you want rescuing from the wreck.
804.
▲
by
dTal
3y ago
It should have been just endless different ways of serving Taystee Wheat
805.
▲
by
dTal
3y ago
>Even 150 years ago, if you wanted to communicate with anyone outside your town, you had to pay a horse based courier what was most likely a very expensive fee. Last month was the 150th anniversary of the introduction of the first govern
806.
▲
by
dTal
3y ago
I do think we can do much better than GPT4. Size isn't everything. Small models can outperform large models, when trained and finetuned in the right way. And transformers are hardly the be-all and end-all of language AI either - there&
807.
▲
by
dTal
3y ago
No doubt. I don't think anyone's arguing that LLMs have richer, deeper understanding of anything just yet. On the other hand I also don't think it would prove much to vaguely connect a language model to a robot arm and then
808.
▲
by
dTal
3y ago
There's nothing qualitatively less "in the world" about a language model than a human. Yes, a human has more senses, and is doubtless exposed to huge categories of training data that a language model doesn't have access
809.
▲
by
dTal
3y ago
They "simplify" the training data, which they are vastly smaller than. LLMs are like compression algorithms. You could imagine feeding the training data back in, letting it guess the next token, and entropy coding the residual - t
810.
▲
by
dTal
3y ago
Thanks for the tip about exllama, I've been on the lookout for a readable python implementation to play with that is also fast and has support for quantized datasets.
More ›