gfodor

50 ms·

631.
by gfodor 5y ago
True - however the thrust of the arguments for opening Twitter seem to be focused on opening up the code - so the “code without data” scenario is the one worth honing in on.
632.
by gfodor 5y ago
Recommender systems like Twitter are as much about data as they are about code. Without the dataset and the derived statistics and models that are used for ranking and recall, the code is not going to help much with transparency.
633.
by gfodor 5y ago
Yep, that’s my guess.
634.
by gfodor 5y ago
I'd imagine the success rate isn't apples to apples - the real measure is "time, energy, and manpower expenditure needed per generated protein"
635.
by gfodor 5y ago
This is a fascinating question, I'd imagine many other people smarter than me have thought about it. Specifically in the domain of the visual arts.
636.
by gfodor 5y ago
Greedy corporations, billionaires, and leftists all pose existential levels of risk to the degree they influence the development of AI.
637.
by gfodor 5y ago
When I say genuinely novel I mean genuinely novel. A trained artist can take a prompt from a person and integrate all they know from art and make something that fundamentally advances both. DALL-E appears capable of this form of creativity.
638.
by gfodor 5y ago
This was obviously the case for previous models. DALL-E 2 seems different - I’ve seen a few outputs that seem like genuinely novel creative artistic works.
639.
by gfodor 5y ago
That’s just begging the question. We don’t know if DALL-E is creative in ways isomorphic to humans. It might be. How would one go about structuring a hypothesis? Have you read the original DALL-E paper? Worth considering: https:/&#x2F
640.
by gfodor 5y ago
How many hours have you spent working with it so far? You obviously have a lot of experience with it, so I’m wondering how you drew this conclusion.
641.
by gfodor 5y ago
AlphaFold
642.
by gfodor 5y ago
Yeah plus these haven’t been trained on multiple epochs and DeepMind just released Chinchilla which implies we don’t need to scale up as much as we have been wrt parameters. Also last month Microsoft published a paper demonstrating deep tra
643.
by gfodor 5y ago
Kurzweil was too conservative. This is the last decade of humanity, as we understand it today.
644.
by gfodor 5y ago
More likely we are at the dawn of something we can’t yet understand.
645.
by gfodor 5y ago
would you mind sharing some references (or even just googleable terms) for this process of fine tuning?
646.
by gfodor 5y ago
Canada literally just froze bank accounts without due process, and the West did the same effective thing at a country-level scale against Russia. I don't blame people who were skeptical of the ideas in the beginning, but as BTC sits at
647.
by gfodor 5y ago
No, Hacker News has been consistently wrong on crypto for a decade, and will continue to serve up these kinds of one dimensional narratives that are wildly detached from reality. The reality is that fiat is in trouble, and crypto is poised
648.
by gfodor 5y ago
I’m pretty jacked into Twitter, but there’s still a higher order bit that seems to be related to people reacting to things like this with broad, sweeping outrage. A lot of people on Twitter either just roll their eyes or think this is funny
649.
by gfodor 5y ago
He’s a great troll, as demonstrated by the reactions here. Why do people get so bent out of shape about tweets? I’ll never understand it.
650.
by gfodor 5y ago
My point isn’t about style but the content and the artistic statement of the content. Both have deep possible interpretations, and if they were drawn by a human artist could motivate a ton of analysis.
651.
by gfodor 5y ago
Based on what I have seen DALL-E 2 does seem to be demonstrating something very close, if not entirely mappable to, human creativity when it comes to visual creation. There are several examples where it makes connections that are both highl
652.
by gfodor 5y ago
You don’t even know what I believe, but one thing is clear: you also haven’t used it yet, and are far more certain of its capabilities than I am. (I have, incidentally, had two of my personal requests generated by the kind folks at OpenAI,
653.
by gfodor 5y ago
A proposition for your consideration: what if you’re wrong?
654.
by gfodor 5y ago
DALL-E 2 isn't the first superhuman AI, but it is the first capable of teaching the whole world of just what that means for all of us.
655.
by gfodor 5y ago
I'm gonna go with unlocked on this one.
656.
by gfodor 5y ago
That's fair. The logic is that elections which do not have any secret ballots, do not have any ballots which go through a secure chain of custody, and cannot be audited are fundamentally election systems that are being run on an honor
657.
by gfodor 5y ago
It’s a horrible idea from first principles, just because people from Oregon were comfortable with the state abandoning the secret ballot and chain of custody in elections doesn’t make it any more sane now that it’s been trust upon the rest
658.
by gfodor 5y ago
The problem, as I stated, is not postal voting. It is mass postal voting. Typically mail in voting is opt-in. Relatively few ballots are printed, and the recipient is engaged to ensure the process completes safely. The blast radius of the
659.
by gfodor 5y ago
Yes, in exchange for removing the guarantees of the secret ballot and a secure chain of custody for ballots. Which, at scale, is a disaster - and basically eradicates trust in elections. The mass mail in voting thing was supposed to be due
660.
by gfodor 5y ago
What about dismissing them after we understand history? Or is the idea we just simply can’t do something once people decide it is (often vaguely) refuted by history?

More ›