fyredge

5 ms·

1.
by fyredge 3d ago
Enron's mistake was doing it too early. Nvidia investing in AI labs to use their own chips falls under anti trust, but I don't think the latest administration is interested in investigating it.
2.
by fyredge 4d ago
Even more reason to show that Canada is making a strategic decision. Both of them are so coupled to each other that they can't afford to hurt the other without hurting themselves. But now that the US has shown it's willingness for
3.
by fyredge 4d ago
It's a great discussion, but not one that I think applies to AI startups. Valve's business model takes the traditional path where profit is generated by selling a product and growing from there. Modern Startups now take large sums
4.
by fyredge 5d ago
I think these two concepts go hand in hand. Perceived cancel culture is the thing that pushed some people to disregard the concerns of others entirely and live without shame. Another example is the concept of involution and lying flat in Ch
5.
by fyredge 6d ago
Points 6 and 7 shows how well the students are intuiting LLMs. I think it ties very well into why they think progress from chatgpt 2 to chatgpt 5 was not a drastic improvement. After all, the first car Ford mass produced and current superca
6.
by fyredge 6d ago
Oh, that's not what I meant. It's the person who shared the content publicly that needs to be deanonymized, not the author. Think of it like advertising on billboards. When you partake in an activity with explicit intent to persua
7.
by fyredge 6d ago
PornHub is a great example! It is a public platform and thus posters have to be deanonymized. For one, this would provide traceability to people who revenge post intimate content. It would not block anonymous viewers. While that means minor
8.
by fyredge 6d ago
I completely understand and actually agree with your stance. Consider this, the personal blog we love reading can remain anonymous, because it is private in the sense that it is ostensibly the writings of a single person, it is not a platfo
9.
by fyredge 6d ago
I draw the line on discoverability. Reddit? Public. Hacker news? Public. 4chan? Public. Facebook and Instagram are also public with a slight caveat, private accounts will be considered private, otherwise public. Even ye old public forums ar
10.
by fyredge 6d ago
But why not? I want to make the distinction on public and private speech. In the physical world, no one should ever have their private speech scrutinized, that's why there are laws banning covert recording. But in public, speech is ver
11.
by fyredge 12d ago
You see funding of chatgpt as a panacea for progress. I see funding of chatgpt as one of small part of a history where governments and industry fund basic science and moonshot programs, not to generate revenue, but to explore what is possib
12.
by fyredge 12d ago
Like funding education. Let's build up human intelligence instead, they seem to have made great breakthroughs in every single field! The US doesn't pay too much to healthcare, they pay too much to health insurance. Too much for to
13.
by fyredge 13d ago
Google was not helpful to me, but what exactly is the deal being discussed? Is NYT going to allow readers to place kalshi bets on their website / app? That would be even more damaging to young men IMO.
14.
by fyredge 13d ago
> UK politicians had pushed for controls covering such features, either through the OSA or separate legislation, but none has materialized. So it's not that online safety act didn't work, but rather that it was not enforced? I&
15.
by fyredge 13d ago
From your comment, I noticed a sort of pattern that is often seen when discussing LLMs. That they are these amazing things that can run so fast they trip themselves in their attempts at achieving a task. So we resort to refining the models,
16.
by fyredge 14d ago
Don't worry about Jesus, he's fine. I'm the same as you sans 10 years of working experience, with similar experiences using LLMs. The distinction I would like to make is that it's not the models that are improving, but r
17.
by fyredge 14d ago
Agreed, though I do think that LLMs are still more similar than we think. Sometime after the first release, AI labs found that coding sat in the niche space of lots of easily digestible data and fast feedback from error messages and compile
18.
by fyredge 14d ago
I'll paste my response to your pre-edited questions: I think atomic weapons peaked during the trinity test. The subsequent creation of bigger explosions by packing more fissile material does not meaningfully improve the technology. Sin
19.
by fyredge 15d ago
Looking at the refutations of Zitrons predictions in TFA, it boils down to two categories: 1. Zitron claims model capability has peaked 2. Zitron claims AI lab growth (user and revenue) has stalled. In the first case, TFA refutes by claimin
20.
by fyredge 15d ago
Well, depending on how you define success, I'd say Trump is successful getting elected as POTUS twice. He sure is a brilliant man to be revered, and the people complaining about his allegations are surely losers... (sarcasm to be certa
21.
by fyredge 16d ago
Very much in agreement. It's the reason successful states have good social security, which provide opportunities. Can't cultivate an Einstein if he can't choose to learn Physics over Farming. The scientists we learn about in
22.
by fyredge 16d ago
Agency. Humans don't seem to need a markdown to decide what to do next. Without agency, it's all tools. I guess that's why a 'tool' is sometimes used derogatorily.
23.
by fyredge 16d ago
This is not the way. You can't legislate away gray area content to children, what you can legislate is algorithmic recommendation of such content. Regulate algorithmic recommendation, revert back to chronological timeline of people &#x
24.
by fyredge 23d ago
It's... not? He's describing how to cultivate founders in university. I'm saying that it's not what a university should be focused on. A startup is not a technical project, it's an organisational endeavour. They don
25.
by fyredge 23d ago
It depends on the book. This style of reading is suitable for textbooks and journal articles, but does not work for fiction, where reading order can be quite important.
26.
by fyredge 23d ago
I disagree with the underlying assumption that it's the universities responsibility to cultivate founders. Universities were initially built to further the knowledge of specific fields, for the sake of its own advancement, producing ph
27.
by fyredge 23d ago
Perhaps so. It wouldn't surprise me if we were to extend an old phrase to "we are what we read". I have seen far too many LLM related discussions break down into politics, philosophy and one too many anecdotes that amount to
28.
by fyredge 23d ago
I find it very interesting that in a website that is ostensibly made up of the curious minded, the top level responses are everything but. From accusations of religious rhetoric, to outright dismissing the argument because they are 'ol
29.
by fyredge 25d ago
Technically, it's still slop. The first example is just slop that you recognise as slop. For reference, see gell-mann amnesia.
30.
by fyredge 27d ago
I will stand by my view that AI cannot be "conscious" until model.train() and model.eval() is combined into one. Going further, it will only manifest when the underlying model, not just the weights, but the layers and the nodes an

More ›