xpl

19 ms·

181.
by xpl 3y ago
> What if Neuralink actually happens, and Musk suddenly has chips in the heads of real people? Yeah, that actually sounds like a more dangerous threat than AGI to me. Maybe all that AGI alarmism from Musk is just to divert the public a
182.
by xpl 3y ago
> but the toy networks we have today don't have nearly the right topology to achieve such states. But why do you think it needs some "right topology"? Why can't "qualia-generating" computation be simple? I
183.
by xpl 3y ago
Is the dataset open source?
184.
Musk considered pulling plug on Ukraine’s Starlink after ‘great talk with Putin’(telegraph.co.uk)
28 points by xpl 3y ago|22 comments
185.
by xpl 3y ago
And yeah... it is interesting that GPT-generated agents sometimes "realize" that they could be indeed GPT-generated. I once saw a strange dream where my "character" realized that he was only a product of my imagination —
186.
by xpl 3y ago
It easily could be. And according to the simulation hypothesis (by Nick Bostrom), it is the most probable thing actually. P.S. There is also a hilarious Rick and Morty episode about a car battery universe: https://rickandmorty.fa
187.
by xpl 3y ago
> If you had a cron script that auto-generated new content every day Thank you for the feedback! I thought about that initially, but turned out that generated jokes become a bit repetitive after a while and the project actually require
188.
by xpl 3y ago
A good idea, did it.
189.
by xpl 3y ago
Check out my recent full-GPT-4 generated HN parody as well: https://crackernews.github.io (featuring coherent comment threads!) P.S. the prompts and the code I used can be found here: https://github.com/crackern
190.
by xpl 3y ago
I wonder what will happen to the remaining Yandex assets in Russia. Will Putin simply nationalize them? If he does, it's uncertain whether the Russian branch of the company can sustain itself, given a potential mass exodus of talent.
191.
Founder of Tech Giant Yandex Slams ‘Barbaric’ Invasion of Ukraine(cnn.com)
10 points by xpl 3y ago|3 comments
192.
by xpl 3y ago
How's Yandex a "clone" if it was launched a year before Google?
193.
by xpl 3y ago
Nice! Check out this GPT-4 generated HN also: https://crackernews.github.io :)
194.
by xpl 3y ago
AFAIK most (all?) available HTTP3 implementations still don't implement BBR as congestion control? Which often makes it inferior in performance to existing HTTP2 on TCP+BBR.
195.
by xpl 3y ago
I don't get why the hell did it take so long (to even begin training, considering how long the training usually takes). If it started a year ago, Ukraine would already have ready pilots.
196.
by xpl 3y ago
It could be scaled up, as it is a relatively "dumb" technology and everything needed for it could be manufactured within Russia, rendering sanctions ineffective. What you observe is just the current state, as it takes time to rero
197.
by xpl 3y ago
You could argue that it didn't work with North Korea as well (as the regime is still there). But in fact the whole country had been bombed to such condition that it simply had no buildings left, it had been totally flattened, ~300k civ
198.
by xpl 3y ago
Russia has large airpower equipped with heavy guided bombs (think JDAMs), which it only has started to use seriously due to the depletion of UA's long-range anti-air defenses (old Soviet S300 systems). The problem is that Western long-
199.
by xpl 3y ago
This is already AGI, as it is not tailored to a specific purpose (unlike all previous "AI"s were). It can do anything (given multi-modality), it is general . Humans are still better at specific things, but LLMs easily beat humans
200.
by xpl 4y ago
https://en.wikipedia.org/wiki/LG_Prada for example. It looks pretty much like iPhone — no keyboard, just touchscreen. But it did not really "spark joy" (I had one), so it wasn't a big success unlike iPho
201.
by xpl 4y ago
Argh, some articles don't fit in a single LLM response and get cut. Maybe add a fake paywall for a more realistic experience!
202.
by xpl 4y ago
Yes! I've also noticed that GPT-3.5 and GPT-4 have somewhat different senses of humor. GPT-3.5 tends to lean more towards absurdism — sometimes it's so bad that it's actually good, but that's not always the case. So, I u
203.
Show HN: Hacker News from Alternate Dimension (With Everything LLM-Generated)(crackernews.github.io)
21 points by xpl 4y ago|6 comments
204.
by xpl 4y ago
> But it is less costly for the cities to remove the problem It does not remove the problem but would simply transfer it elsewhere. In result, either people become significantly less mobile, or they get more used to cars, creating huge
205.
by xpl 4y ago
> Since it's a rental, nobody wears a helmet and almost everybody drives them irresponsibly In Tel Aviv, all rental scooters come with helmets, and wearing a helmet is actually enforced by the police. Also they are banned from rid
206.
by xpl 4y ago
The docs, for the reference: https://ytsaurus.tech/docs/en/ P.S. I wonder if LLMs could be used to generate docs and comments for big hairy codebases. Seems that the current generation of LLMs lack context to do i
207.
by xpl 4y ago
My guess it could be successfully used for relatively small (terabyte-scale) workloads. It's the same as when people use k8s not utilizing its full capabilities, only to be able to massively scale up when needed.
208.
by xpl 4y ago
It isn't new, in the article they state that it originated back in 2006. It's been in development since then.
209.
by xpl 4y ago
I doubt about "a lot". Also, "already solve" does not mean "solve better" or even "good enough". It would be very interesting to see some in-depth comparisons with already-existing open source technol
210.
by xpl 4y ago
AFAIK it wasn't in the leak (that "YT" platform specifically). Also, open-sourcing proprietary projects of this scale is a pretty hard job and it can't be done quickly — seems they had been doing it for a long time, star

More ›