I don't know where else to write this, but I want to throw the idea out there. I have long wanted to take old broodwar televised matches, many of which are terrible quality 240p, and use machine learning to convert them to into perfect Broodwar Remastered frames. This seems tractable to me because you should be able to map the terrain sets to their remastered equivalents, and the game is just a series of sprites rendered at specific frames. Even if the source quality is terrible, I imagine this is able to be extracted at high quality since you can, eg, set up an automated pipeline which generates training data. Maps from original graphics to remastered, and then again for 240p -> tilemap positions for frame camera center + sprite positions / animation index.
OSN did it on their youtube for their old OSL series, I think it can be done better but I think original still is much better to look at, at a corresponding resolution.
The key expression for playlists or yt browsing is this `AI 업스케일`
Do replay files require consistency or can you spawn new units from the ether as required? I think you could get a system that would get the broad strokes unit placement correct, but having a shot-for-shot perfect replication seems impossible.
> I was wondering how you could possibly sync up something like a Carrier that has multiple fighters.
Same way as everything else - you play out the simulation and see what happens. The behavior of the interceptors is as deterministic as any other unit. (They just happen to not be directly commandable by their owner.)
No, I get that. The pondering was before I knew it was a recording of actions. Once you know that, it is a fully self-contained box with external nudges from the players which get inserted into the simulation.
Not feasible. The VODs are from a spectator's perspective; there's a lot of action happening "off camera" (like players building units) which affects what's seen on camera later, and which would be extremely difficult to reconstruct from what's visible.
I had a similar idea for Super Smash Bros Melee! There is a lot of very low-quality footage out there; surely you could train some sort of model to convert it to Slippi replays. A year or two ago this would have been a grad student project; a year or two from now, it will be a one-shot prompt.
I found the post offputting because he is insinuating a few things about the author which are clearly untrue (eg, unfamiliar with the formal verification landscape) and he's using this untrue speculation as evidence of the dangers of vibe coding. Even the edit where he says swap it out with a hypothetical person who fits the description - it's still leaving in the untrue claims about the author.
It doesn't seem unreasonable to think it mean if I were to make up speculative negative assumptions about you which undermined something you've been working hard on for a long period of time.
I know nothing about the author, I'm just watching this whole spectacle unfold.
My problem with flatpak is that the underlying runtime has a limited lifespan. I used the Sublime Text 3 flatpak for many years, and it was deprecated a few months ago because the runtime it was using was no longer supported. This caused people to update the flatpak to Sublime 4, which is good, but I only have a license for 3. In the end I just downloaded the binary for 3 from their website, and it still just worked. Flatpak still doesn't save you from bitrot.
I have had my account since like 2013. I browse the site and hit like on things, but I only have a handful of posts going back over the years. Occasionally, when I do try to post, it stops me because they think I am a bot. Meanwhile, I see what are clearly and unambigously bots follow me all the time, I see identical messages repeated across obvious botnets. Their bot detection is just very bad.
Also, while we're on the subject, I used to have two factor enabled. After Elon bought the site, they forced me to disable two factor for insane reasons, iirc you needed to pay for a blue checkmark if you wanted two factor. Today for the first time they've started blocking my access to even view the site until I authenticate with a qr code from an android or ios device. The future of the internet is bleak.
Everything is locking down so hard. When someone says they got banned without breaking any rules, I believe them because it happened to me, but also assume there's some reason they appeared out of the ordinary. It's not going to be fair either, like they trust phones more than PCs.
You're comparing the challenge of building game engines in the 90's with the more open-ended AGI research. It's different enough that I don't think it's valid to conclude he's less productive now based on impact over time.
People point to the equivalent API costs to show that they are getting a great deal on the subscription, 10,000 dollars worth of tokens for 200 dollars. I do wonder if it's the other way around though - are the API users simply getting ripped off? I have seen Dario say in multiple interviews that they are profitable on inference, which maybe he was only meaning to refer to API usage, but that's not the impression I got.
It's not a 98% margin loss if your users are unwilling to pay 50 times the cost that they were previously paying, and if they have other options like open source providers. The calculus isn't so simple because some portion of users would switch to API, and so it's about how many would continue using the service rather than leaving for a competitor.
I'm aware they need to recoup the enormous cost of training and data centers, but on a purely inference cost level I'm not convinced that the 200 dollar plans are unprofitable.
> I'm not convinced that the 200 dollar plans are unprofitable.
Especially considering not everyone is tokenmaxxing, and in most parts of the world people take leave and companies do not cut their subscriptions.
I suspect they are priced to have a lifetime average price/token amount that is roughly break-even, or maybe a slight loss leader.
> have seen Dario say in multiple interviews that they are profitable on inference, which maybe he was only meaning to refer to API usage, but that's not the impression I got.
I think he does mean API usage. Don't forget they can (and do) adjust the number of tokens you get on each plan at any time to adjust their margins on those.
That means he knows that is controllable, and it only the underlaying inference that defines the succes or otherwise of the company.
> Especially considering not everyone is tokenmaxxing
Exactly. I have the Claude $100/mo plan, and use it moderately for open source hobby stuff. I still haven't dipped my toes into the Fable pool, but I always use Opus 4.8 on xhigh, and I never hit my limits.
On the other hand, though, there have been times when I've looked at /usage for a long-running session (e.g., 7-10 days, after it's compacted a few times), and it showed I'd used ~$450 worth of tokens just for that session. So I'm clearly getting value for the money here when it comes to the subscription cost. But I still don't hit limits, so...
Yea, it’s like pointing at the cost of renting all individual movies and TV-series at Netflix and concluding that Netflix subsidizes the subscription with tens of thousands of dollars.