The compute cost for DeepSeek V3 was $5.6M. Training costs for traditional US frontier labs model is about $70 to $200M+, with estimates of up to $1B+ to train the next generation models. It's expensive, but it's not even in the ballpark of "There is no viable economic model to do so." Google revenue in Q2 was $1.32 billion *per day*. $1B is trivially less than a day; $5.6M is just 6 minutes of revenue.

The cost problem is in the compute buildout and the subsidized compute to capture users.

> It remains to be seen if even using their models can be profitable,

I presume we can stipulate there exists some tasks that are cheaper to do using an LLM than doing it "by hand". As a such, using an LLM model can be profitable. To me, it is also obvious that the business of serving an LLM model can be profitable. There's large benefits to being able to get tokens quickly, and you can expect people to pay for that, even if we presume they could run the model locally with no maintenance overhead. There's even cases where using the cloud is cheaper than you can serve locally at all - I saw one guy that measured his added cost of electricity when his computer ran a specific LLM, and it was more expensive to run it locally than to buy tokens from the cheapest inference provider for the same model, presumably because they had better hardware and cheaper electricity.

I love a term a friend of mine used to use when talking about AI: "Narrowly superhuman" (before "superintelligence" got to be a thing).

AI is narrowly superhuman. It is able to do specific things that used to be defined as a "hallmark of intelligence" much better than a human. The easiest one? Summing numbers. It's both faster and more precise than a human.

AI gets more and more such areas it can do better than a human.

If we define "Superintelligent" to be "A computer system able to do absolutely everything we can somehow call intelligence better than any human", we will *never* have superintelligence. The things we define as intelligence will be in conflict; we'll land at the system being better at one at the cost of another.

I'm in favor with being careful about this development. But "prohibiting the development of superintelligence" isn't an option. It's already here.

Dividing the yearly US spend on Internet advertising by the number of people in the US that has Internet access gives $900-$970 spend per person per year, or $75-$81 per month.

Please choose the option you want:

1. I want to pay about $75 per month per member in my family into the ad auctions to not see any ads. Ie, keep the infrastructure, but let you opt to pay your share through money rather than attention.
2. I want to prohibit ads, and accept that everybody will have to pay for the sites they want access to, that there will be much fewer services, that there will be much higher transaction costs, and that the poor will have much less ability to access services. (Microtransactions won't solve this, due to the need for trust evaluation of same.)
3. I want some combination of 1 and 2, where some sides of ads are prohibited so it is possible to buy my way out of them at a lower price,. I accept that this will result in the loss of some services and lower access to others for the poor.
4. I want legislation that makes Internet ad blockers work more consistently. I accept that this just shifts the attention cost to other people, and have me getting the benefit without paying the cost.
5. I want to remove Internet ads and subsidize the sites through taxes.
6. I want to increase the requirements of Trust and Safety checks and format restrictions for ads, potentially accepting higher overhead (in what goes to content vs infrastructure) for less annoying/bad ads. I accept that doing this creates significant advantages of scale for ad networks.
7. Other, please describe, including where costs gets shifted with your proposal.

The above are my best attempt at listing possible options and their correct caveats. In most situations, I find ads annoying (who doesn't). Unfortunately, they are the main money highway of Internet content.

DISCLOSURE: I used to work for Google, where I rang infrastructure for stopping bad ads/advertisers and for removing fake clicks. I've also been interested at the use of sub-cent microtransactions to pay for content since the 1990s, though I've become disillusioned as I've learned more about the attention economics involved. Approving individual microtransactions has a higher attention cost than dealing with the ads, and bulking them per-site (through micro-subscriptions) will limit the number of sites that can have content.

OK, which particular aspect of the LLM learning do you think doesn't exist in the human brain? Backprop exists (and the other training methods the brain uses has similar effects), embeddings more or less exists and seems to be very similar to the LLM embeddings (shown by brain activation in various areas linearly matching with embeddings) - see e.g. High-level visual representations in the human brain are aligned with large language models (Adrien Doerig et al, 2025), and speech/text generation by humans has similar error patterns as for LLMs.

They obviously don't learn in the EXACT same way, but there's also clearly a lot of bits that are the same, and I'm fairly sure nobody knows how much is similar yet, particularly not weighted by importance.
 

It will favour zealots in both directions, which should mostly even out assuming zealotry is roughly as common in both directions. Also, I don't know of any less zealot-based approach except forced participation - which seems worse.

Citizen conventions typically consists of randomly selected citizens. This one consists of 150 randomly selected citizens using a statistical procedure (sortition, a variant of stratified sampling) to make it represent the whole.

The FAQ on the website says this:

It also links to the detailed terms and conditions, which I've not read.

From the FAQ:

Does Google own the intellectual property created during the competition?

No. Google is not requiring any IP or licenses be granted except a non-exclusive license to be used only for the purpose of testing the inverter and publicizing the prize. We want entrants to benefit themselves through the advancements they make in order to help grow an advanced power electronics ecosystem.

It could be that he's gotten his PhDs the old-school way: Doing his research in the field and then presenting to a committee, not guided research. (I don't care enough to look it up, I just want to make sure that people that read the thread know that this option exists.)

... but I'm not sure whether it counts. I was the last regular employee left when a company was being shut down, and part owner of that company.

You got an extra zero in there, right? As in 7 years sounds about right?

I know some authors protest that seven years is too long, and the majority of income is made in the first three years and after five it would be advantageous to have the works available in the public domain (but the publishers don't want the competition from previously released works), but I think that varies from author to author, so doing a compromise of seven seems reasonable - we can experiment with shortening it further after having seen what happen when we cut it to seven.

Anwar al-Awlaki - intentionally executed without trial.
Samir Khan - unintentionally executed as part of the execution of Anwar al-Awlaki.
Jude Kenan Mohammad - intentionally executed. Had previously been convicted of terrorism conspiracy, but not to a sufficient degree to actually be imprisoned.
Abdulrahman al-Awlaki - 16 year old with no personal involvement in terrorism, but who had a father (Anwar al-Awlaki) who was involved. Robert Gibbs, former White House press secretary, stated that he "should have had a more responsible father." Unclear at what level the execution was a mistake.

Three of the four are arguably "bad guys" - but they should still have gotten a proper trial, so we could determine if they are. The last one doesn't even seem to be a bad guy, just somebody that happened to be born to an unfortunate father.

That's a common meme for people that don't know the psychology of suicide. Suicide generally takes a fair bit of strength of character. The problem is not being "a pussy" or "taking a pussy way out" - it's an unfortunate evaluation, often from wrong premises. A common one is that a bunch of psychological or life situation issues are permanent/long term, while they are usually possible to change. Another is overestimating how much the person is a burden to other people in their lives, and wanting to lighten their load.

Yes, I have. It's embarrassingly low. A little less than what Greece gives, about half of what Germany gives, about 1/5th of what Sweden gives.

There's some stats over at http://www.statisticbrain.com/countries-that-give-the-most-in-foreign-aid-statistics/

The US has a lot of good points. Foreign aid isn't one of them, and neither is consumption patterns.

(Oh, and I live in the US and am originally from Norway, if that makes a difference.)

The transaction costs for paying bitcoin are fairly high. I'd need to find somewhere to buy bitcoin (Mt Gox, I presume), find out how to pay with it, probably install some software, etc. So that's a hassle even though I know a bunch about Bitcoin, as I haven't used them before - and hassle is a major part of transaction costs.

If you're looking for more technical transaction costs, they're about $6.59 per transaction at the moment, paid by the entire network through expansion of the monetary base (using today's block payment of 25 bitcoins per block, the rate of one block every ten minutes, yesterday's transaction count of 51923, and this moment's cost of bitcoins of $95) . That transaction cost will drop if the hashing power compared to buying activity goes down.