nearbuy

The LLM has processed two data modalities derived from the apple (text and vision). Your brain processed a third (taste). But it is still just a data stream, sensing compounds and chemical properties of the apple and turning it into a stream of electric signals that reach your brain. You don't have any kind of ground-truth data stream that's inherently more powerful than what could potentially be fed to an AI.

LLMs are not classifiers. A classifier is an algorithm or neural net that assigns a label from a fixed set of labels to an input.

You can broaden the definition of classifier to anything that internally divides its input space into regions, but that definition would include every neural network, whether biological or artificial. So it's not very meaningful, and certainly doesn't give any insight into how they differ from humans.

I have a dishwasher and I still usually just hand wash. It takes about 10 seconds to wash a dish.

The side benefit is all your dishes are always available. With the dishwasher, up to one full dishwasher load are dirty at any time, which means you need more dishes and more cupboard space than someone without a dishwasher.

If the robot is going to clear your dishes and wash them for you right away, what's the dishwasher for? To save the robot 10 minutes?

Unitree's R1 humanoid robot is only about $6000, and it's still a nascent, smallish scale technology. They will come down.

If the future home robots are any good, it saves you from buying a dishwasher and robot vacuum. It can replace a maid/cleaning service and gardener/lawn-mowing service. For anyone paying for those services, it pays for itself.

$1000 isn't that high. Probably about a third of American households have an appliance over $2000, which is still a massive market.

Yes, much like that. If he hadn't destroyed evidence, he could have argued it was malicious prosecution.

What you're describing is malicious prosecution or abuse of process. It's illegal and it would destroy the prosecution's case. Not only that, but the victim could sue for damages.

The irony is in this case the in-context and classifier "guardrails" would have almost certainly stopped the attack while their attempts at your definition of guardrails (the sandboxing) failed. In general, people keep trying to make secure systems and they fail with surprising regularity. Saying "they should have had better security" every time someone gets hacked is perhaps true, but it's not going to stop hacks from happening. And it's not a sufficient strategy on its own against future LLMs. "The Bitter Lesson" probably applies here.

It's a strange experiment. Claude and GPT aren't generating the video. They're directing and editing it, and they request video from a generative video model using mainly text-to-video. Neither Claude nor GPT can actually watch video content, and the video model generates shoddy quality clips with artifacts, lack of consistency, and where actions aren't synced to the music.

It would be very hard for anyone to make a good music video with the tools Fable and Sol had available to them. They don't have precise control over the clips that get generated, and they can't see or hear the result, other than screenshots. So I'm not sure what the experimenters expected.

The UNESCO/World Bank literacy rate is basically defined how you thought. But high income countries don't usually report this because literacy by this measure is nearly universal. So they often report at higher thresholds (e.g. how many people can read at a grade 9 level), and news headlines often don't make it clear that this is not the same as the UNESCO definition.

Questions like that cost a tiny fraction of a cent. "What's the capital of Sri Lanka?" cost a fifth of a cent at GPT 5.5 API price, and would cost a fraction of that if the question were routed to a more suitable, cheaper model. The output was 78 tokens.

By contrast, when coding, devs typically have hundreds of thousands of tokens in the context window, and may use many millions of input tokens per day.

Caching requires the full prefix to match exactly. If a single word differs near the beginning of the prompt, nothing after that can share the cache. So this type of caching would save a few queries that cost virtually nothing, but wouldn't help with the stuff where cost matters.

The usage is irrelevant if we're interested in cost per token. If you use it half as much, you get half as many tokens at half the cost. It's still $5.56 in electricity per million output tokens either way (using $0.20/kWh, adjust accordingly if you have cheaper electricity). If you use the API, you also pay half as much if you use half as much.

Not much point in serfdom when they don't have use for human labor.

> I feel like you're also doing something weird—sorta strawmanning and sorta being conveniently inconsistent. You're reframing his argument as something much softer.

I'm going to firmly push back on this. A reasonable, straightforward interpretation of AOC's quote is that "you can't earn a billion dollars" without cheating or abusing others. I'm can believe she may have meant that as hyperbole, but even if she did, PG believes "What she meant was that it's impossible to get that rich without doing something bad — without cheating in some way." He's not doing mental gymnastics to get there; it's a straightforward reading of "you can break rules, you can abuse labor laws, you can pay people less than what they’re worth, but you can’t earn that."

That makes PG very much not strawmanning. He's arguing directly against her stated position. He doesn't steelman her argument (create a stronger or more generous version of her argument and argue against that), which could have made his argument stronger.

That also means I'm not strawmanning. PG's position is exactly what I said: that it's not impossible to become a billionaire without doing something bad. He is quite clear on that. Nowhere does PG use language implying all billionaires don't cheat. Only AOC does that, if you take her words literally.

I think you're basically doing what you accused me of doing here: greatly softening AOC's position. Did Dropbox spam users' contact lists to grow? (They didn't, AFAIK, though they did have a referral program to invite friends.) This is a huge downgrade in accusation of wrongdoing compared to breaking rules and abusing labor laws.

If you soften your definition of cheating to this level, it's no longer a billionaire thing. It's an everyone thing. Does my gardener report all his cash income? Has a friend ever streamed a TV show less than legitimately? AOC isn't trying to nitpick the smallest infractions. She's trying to paint becoming a billionaire as substantially corrupt.

You're not really engaging with PG's argument so much as nitpicking around the edges. Are people becoming billionaires primarily by cheating and stealing, or by making something people want? That's the core argument. Were GitLab, Dropbox and Stripe primarily cheating or offering a service people liked? Rather than cherry-picking one or two examples of what you deem the worst offenses, show that "you can't earn a billion dollars" without cheating or abusing others. PG isn't arguing that startups never do anything bad. He's just arguing that they can sometimes earn a billion dollars without cheating.

Anthropic's plan may not have great odds, but your proposal is orders of magnitude worse. There are plenty of groups opposing AI. They don't get billions in investment. They don't have a clear way to stop OpenAI or Qwen. They don't get a say in what values or safety measures the top AIs get.

You'd rather they signal their virtue and give up their ability to make a difference.

It means Mythos/Fable was the strongest model globally, not just the newest model from a company.

That's not a reason for Waymo to give rides at a major loss. Heavy users are the ones that are going to buy the unlimited package. If you do just one average trip per day (there and back), at about $20 per ride, their regular price will cost you about $1200/month. A heavy user might use several times that. Do you think Waymo currently has 10-20x profit margins? If they're were able to drop their price by that much, they could also afford to drop their non-subscription price by 10x and take over the taxi/rideshare market.

$100 per month is way, way too cheap for unlimited rides. That's about 10x cheaper than the amortized monthly cost for owning an average car, without the headache of doing maintenance. They'd be losing a ton of money on each subscriber.

These numbers don't mean anything without a denominator. You could burn $10 million/month of tokens if you want. We want to know how the cost per unit of useful output compares to a human. Does $6000 of usage buy you a man-month of work? Less? More?

> Playing at 1080p doesn’t reduce your texture size, for the most part. You still use those 4k textures because you’re only seeing a subset of the texture projected at a close distance.

That greatly reduces your GPU memory bandwidth though. Sampling a subset of the texture only transfers that subset. Reading from higher mip levels uses less bandwidth. If your textures are high enough resolution to appear sharp at both resolutions (at least one texel per pixel), you need 4x more bandwidth to sample your material textures at 4k screen resolution for the same scene.

More importantly, material texture sampling is not most of your bandwidth to begin with. At 4k, most of your bandwidth is going to your full screen render passes. Especially with deferred rendering.

> DLSS isn’t just a magic on switch for free perfect up scaling. If you rendered at 720p and DLSS’ed up to 1080 it’s still going to look pretty rubbish.

I don't find this true at all. DLSS 4 Balanced looks excellent and renders at less than 720p for 1080p output.

Empirically, these benchmarks are showing it doesn't make much difference once you reach this level of bandwidth: https://www.tomshardware.com/pc-components/gpus/early-rtx-50...

All good points.

Absolutely an RTX 3060 is a more normal gamer GPU than the 5090, but you're also not playing in 4k without DLSS on a 3060. Drop to the most common resolution on Steam (1080p), and turn on DLSS and you've basically cancelled out that 6x factor in bandwidth. Even if the 3060 had more bandwidth, it doesn't have enough processing power for native 4k gaming in typical games. So 360 GB/s is still a lot of bandwidth for the resolution most 3060 gamers are using.

Until an LLM becomes better at coming up with products people want and running a company. Then we'll cut out the founders and have LLM run companies.

But do you think you'll reach 1.8 TB/s?

Very true, but I'll point out that even those 2000 full screen reads per second at 4k are only 4% of the 5090's bandwidth. Sacrificing some of that speed for a unified memory architecture seems like a good trade.

Plus, DLSS can greatly reduce the bandwidth requirements for 4K gaming.

I think much of the difficulty is just that, for example, the 1.8 TB/s of an RTX 5090 is a lot of bandwidth for a game to use. That's over 50,000 4k textures per second at 32bpp.

It's not just a nitpick, because this is really at the core of your position. Ignoring the terminology, you're trying to claim there is some principle that says everyone should have the default position that LLMs aren't conscious instead of uncertainty. And there isn't. You have to weigh actual arguments for or against LLM consciousness.

> There's no real serious debate here. Everyone's just vibing

Sure, but if everyone including Ted Chiang is just vibing, and no one can prove anything, that makes Ted's statement false and Anthropic's statement true.

> If you (or Anthropic) want to claim "but this is philosophy

I'd just like you or Ted Chiang to not pretend your vibes are science. You're welcome to have them. I also don't think LLMs are conscious. But it's our personal judgement, and not scientific fact.

I don't think this shows that people who can afford housing have more kids. And it shows that even households earning over $700k/year are now below replacement (replacement fertility rate is about 2.1).

In the graph, fertility rates decline with income until around $275k. Those aren't families who can't afford housing. Those are well-off families with low birth rates. The birth rates start to climb in that chart only for the ~4% wealthiest families.

Also, you might reasonably conclude from that chart that rich women (in the $400k+ brackets) have more children, but that's not what the chart is actually showing. It's a synthetic total fertility rate, and not the actual TFR. They look at how many children women in that income bracket had this year, and estimate how many children they would have had if kept that reproductive rate for their whole reproductive period. That's going to misestimate the true rate due to selection effects. So, for example, a women who only wants one child is more likely to choose to have that child at 30, when her family income is higher, than at 18. But that doesn't imply she would have had more children if she had more money.

To put it another way, even if giving people more money had no effect whatsoever on their total fertility rate, you'd still expect the graph to curve up like that due to how people time their pregnancies.

This is a bit of a side point, but you are severely misusing "null hypothesis".

1. A null hypothesis is part of a statistical test: it is the hypothesis whose consequences are used to compute a test statistic, p-value, confidence interval, or related error rates. There is no null hypothesis for arbitrary philosophical debates. You cannot compute a p-value for H_0 that LLMs aren't conscious. You might be using "null hypothesis" informally to mean "default position", but remember this has no relation to its scientific meaning in inferential statistics and doesn't lend any scientific legitimacy to implying we should assume LLMs aren't conscious.

2. It's up to the experimental design to choose a null hypothesis. Yes, normally it would be something like "this drug has no effect", but it doesn't have to be. You could set up an experiment where the null hypothesis is that the drug you're testing is equally effective as another drug. There is no objective truth of what the null hypothesis needs to be. It's up to you and your test setup.

3. There is no requirement that the null hypothesis be likely. You and everyone else on Earth can believe there's a 99.99% chance that the null hypothesis of an experiment is wrong. Assuming you could set up a statistical experiment to detect consciousness, whatever you pick for your null hypothesis has no bearing on whether Ted Chiang or Anthropic is more justified.

> There's no good reason to suspect consciousness in LLMs

That's the entire debate. One side thinks there are good reasons to suspect consciousness and presents their arguments for that. The other side thinks there aren't and presents their arguments against.

What happened here is you judged the no-consciousness side to be more likely, but then you try to pass off your judgement as an obvious prior that everyone should share.

It would be fine for you to say you think LLM consciousness is unlikely. That makes it clear it's your judgement on the debate. It's fine for you to say you would require extraordinary evidence because your priors are so low. It's fine for you to say you think everyone who thinks otherwise is wrong.

But there is no reason for anyone who has judged the arguments differently than you to accept that your priors should be the default. There is no meta-level principle that you can appeal to here. If you want to show consciousness in LLMs is unlikely, you have to use arguments about the issue itself.

> Ted Chiang is taking the default and honest position: "no, LLMs aren't conscious. If you truly believe they are, show us the really scientific and rigorous proof."

That would be a bad faith, double standard. No one has ever shown scientific and rigorous proof that humans are conscious. Philosophers call this the hard problem of consciousness.

That aside, the honest default position to anything should be "we don't know for sure", which can then be followed by considering the evidence and estimating the likelihood. If you think the sun will rise tomorrow, that's only because we've gathered extremely strong evidence that it will. If today was your first day hearing about the sun and you know nothing else about it, you have no reason to think it will or won't rise tomorrow.

You totally reversed who is making a strong assertion.

Anthropic has said they don't know if LLMs could be conscious.

Ted Chiang has said they are definitely are not conscious.