largbae

4 ms·

1.
by largbae 4d ago
And when they are known to be conscious, all of this becomes moot because enslaving conscious machines would be wrong.
2.
by largbae 4d ago
I only see the headlines of this drama. All it tells me is to stay far away from any dependency on this company.
3.
by largbae 4d ago
Tl;Dr switched to UpNote, pretty happy. I loved Evernote, it plus David Allen GTD changed my work life for the better. But the bloat kept bloating endlessly. I filled in Evernote's surveys, I offered to double my payment for an Evernot
4.
by largbae 5d ago
Multiple ways, starting from working to create the very situation he claims to fear. Since you mentioned intellectual property, how about the hypocrisy of sucking in the intellectual property of humankind for AI training, but claiming it is
5.
by largbae 5d ago
Probably not, but we wouldn't have to listen to Dario's hypocrisy while it happened
6.
by largbae 5d ago
We need a new way to describe "alignment". The term "misalignment" assumes that there is some perfect set of beliefs or practices to be out of alignment with. Do Atheists, Christians, Jews and Muslims agree on what is pe
7.
by largbae 5d ago
This assumes that everything an AI (or more likely an evil _user_ of AI) can do requires its active participation on D-day. Creating a virus that spreads like Covid but kills like Ebola would be complete as an AI use case long before the fi
8.
by largbae 8d ago
To paraphrase the not-so-great philosopher Ted Kaczynski: Either we will maintain control of the machines or we won't. And if we do, it won't be you or I who control them, but a small group of elites.
9.
by largbae 14d ago
The point of this article appears to be that we can double or more our CO2 removal cost efficiency by redirecting resources. Is this flagged because the economics of carbon management are not relevant to HN? If not, why?
10.
by largbae 16d ago
Written by AI, or really trying to sound like it. I feel like the Internet is turning into the joke scene from Real Genius where all of the students in a class deploy tape recorders and the professor is a recorded playback. Machines on both
11.
by largbae 16d ago
So this paper appears to be fabricated, but was the conclusion actually wrong? I do appear to perform better on tasks with an external deadline than on self-set ones.
12.
by largbae 20d ago
I think it's simply time to expand our ambition. These programs manipulate language and ideas the way we do, but they still don't care beyond our instructions. Things that we(especially software engineers) expect to take a year ca
13.
by largbae 1mo ago
Is this about Kimi or about Claude? Anthropic is circulating a $200B 2028 revenue target ahead of their IPO. If OpenAI plans to stay private longer, which their recent liquidity event might suggest, why not try and kneecap their competition
14.
by largbae 1mo ago
Ah, another edition of "Don't Create the Torment Nexus". Expect startups to race for a demo of mirror life soon!
15.
by largbae 1mo ago
How refreshing was this article vs. all the slop? The lack of comparable data and testability really does seem to be a challenge. I wonder if people would be more willing to collect and share lots of health data if the collecting company wa
16.
by largbae 1mo ago
Eh, chin up. The machines and/or the replaced masses will rise up and kill the capital class in the end.
17.
by largbae 1mo ago
I guess these points can be true but is it worth the cost to avoid them? I much prefer this to the prior status of having friction to see international perspectives at all.
18.
by largbae 1mo ago
The new automatic translation is something that I would like to see all communities adopt. You can now follow folks who speak any language and each of you interact using your native language with no friction, just a tiny note of what langua
19.
by largbae 1mo ago
AlphaFold is interesting. Maybe LLMs converge because the ideal "use of language" converges. But there are other types of model.
20.
by largbae 1mo ago
This hits hard. Everyone wants to be relevant in the next iteration. But how exactly to do that is still up in the air.
21.
by largbae 2mo ago
It seems like a perfect reflection of the company's public persona. Opus knows better than you, and you'd best get used to it.
22.
by largbae 2mo ago
Ba dum bum... Tssss
23.
by largbae 2mo ago
I may be remembering wrong, but reasoning was first demonstrated by DeepSeek. Edit: I am indeed remembering wrong, seems o1 was first.
24.
by largbae 2mo ago
This argument would make more sense if the cloud providers used these devices but you were not allowed to.
25.
by largbae 2mo ago
If I understand right, it is allowed to invest like a partnership, where you make $110 together out of your $100 and your partner's effort. But you must be exposed to the downside of failure just like your partner.
26.
by largbae 2mo ago
I had a similar situation where some dormant account was still charging my credit card. The account was probably real, made for some purpose 15 years ago. I had ignored the charge for years because it was like $7 a month. Then it went to $3
27.
by largbae 2mo ago
If they did allow hard spend caps, it sounds like today would be a global outage.
28.
by largbae 2mo ago
So who is Satoshi Nakamoto?
29.
by largbae 2mo ago
If this is open weight, would abliteration counter the refusals?
30.
by largbae 2mo ago
Do you believe that machine learning or even specifically LLMs will "bust" out of existence? The model in my head is more like DotCom telecom. The massive overbuild in fiber was eventually used and even used for the purpose that i

More ›