The cost can be significantly reduced immediately and drastically if OpenAI or Anthropic were to choose to do so.
By simply stopping the training of new models, profitability can be achieved on the same day.
With the existing models, we have already substantial use cases, and there are numerous unexplored improvements beyond the LLM, tailored specifically to the use case.
This only works if all the AI companies collude to stop training at the same time, since the company that trains the last model will have a massive market advantage. That not only seems extremely unlikely but is almost certainly illegal.
Current frontier models are not good enough because they still suffer from major hallucinations, sycophancy, and context drift. So there has to be at least (and I have no reason to believe it will be the last, GPT-5 demonstrates that the transformer architectures are hitting diminishing returns) one more training cycle.
US universities do a lot of research on Palestine and Israel's occupation and apartheid there. In addition, universities are naturally liberal and are targets of the right wing anti-education movement.
And you're offering nothing to back up your claim. Literally everything written in the article is backed up by sources such as New York Times, which in turn quotes experts in the field.
- Is it not true that the existence of a link between high cholesterol and heart disease is only a hypotheses that originated in the 1950s, not a scientifically proven fact?
- Is it not true that the FDA generally has not required drug companies to prove that cholesterol medicines (such as statins) actually reduce heart attacks before approval?
- Is it not true that lowering cholesterol by different means (i.e., other than statins) is not beneficial, and does that not mean cholesterol is not the villain it is made out to be?
- Is it not true that the only large clinical trial funded by the government (rather than drug companies) found no statistically significant benefit at all?
After seeing this post and the discussions in the comments, I created a poll to gauge how many hours the HN crowd typically works. Curious about the results.
London isn’t an example of cultural erosion—the pie has simply grown. The same applies to New York. Diverse cultures enrich London; they add to it rather than diminish it.
Absolutely absurd. Why is it people feel the need to make out that western cities are somehow magically immune from the exact same detrimental effects that they happily accept about everywhere else?
"Enrichment" is a buzzword for the insulated elite happy that they have new things on their lunch menu, somehow ignoring all the negatives that come with it.
I can't help observing that Britain established a global empire (so vast in scope that they could truthfully say the sun never set upon it, until quite recently) and made it clear to the countries they colonized that Britain, and particularly England, was the center of civilization. We're surrounded by the artefacts of this empire, from Imperial measurements in the USA to Greenwich Mean time being the default timezone to which all others are calibrated.
If you establish a vast trading empire, and tell the often surprised new inhabitants of it that the empire requires their spices/ silks/ slaves, can you really be surprised that the more enterprising and adventurous colonised people gravitate toward the point of origin? Is it some sort of mystery that there should be more people from Algeria and Congo in France, or why there are so many Indonesian people in the Netherlands?
I feel a similar perplexity about many people in the USA making loud complaints about cultural adulteration despite a good quarter of the land having previously been part of Mexico and this being reflected in most of the place names (to take but one example). Some commentators object with an absolute straight face to hearing so many people speaking Spanish in cities with names like Los Angeles.
Migration is a policy issue, not some natural force like osmosis.
It's entirely irrelevant why migrants want to come, when you have an existing population that is being harmed by migration policy.
It's simply that the wealthy benifit from increasing labour supply; this is the "enrichment" you get. Not an improvement in culture or conditions, an increase in wealth concentration.
It's entirely reasonable for a citizen in an English speaking nation to complain about other languages becoming common, as it is direct evidence of the migration policy that is doing them harm.
Sauce for the goose is sauce for the gander. I am indifferent to the plight of former colonial nations experiencing the same phenomenon in reverse.
Migration is a policy issue, not some natural force like osmosis.
There's a whole discipline called social physics successfully leveraging models from physics, chemistry, and mathematics to analyze and predict social phenomena. Just as the behavior of flocking birds can be reliably modeled with a few differential equations, so can a great many social changes.
There is a disgusting degree of callousness with this comment.
It's not a "colonial nation" that is being harmed by migration, it's working class people. Meanwhile the same classes that benefited from colonialism now benifit from the wholesale destruction of communities.
Every empirical study I have seen shows that high migration lowers happiness and trust. Does that align with your pseudoscience?
Not as callous as pinning all social problems on migrants who mostly just want economic opportunity, and who generally commit crime at lower levels than the domestic population. Nor is it as callous as shrugging off the social and economic losses inflicted on those countries when they were colonized in the first place, or 'granting' them freedom without compensation for the looting, murder, and de facto enslavement.
working class people
I've got a feeling you're not a socialist, though.
Always amusing that lowering the percentage of natives in a culture is an improvement, "enrichment", "vibrancy", and so on, but only in one particular kind of direction...
And interestingly, the people who utter this ideology are usually from the natives - because fundamentally it's a type of narcissism - a kind of "I'll show you" or "I like you being hurt" that feels pleasurable or gives a sense of superiority.
Can't imagine this ever being tolerated within China and Russia though, perhaps that's why they're the current bogeymen.
By simply stopping the training of new models, profitability can be achieved on the same day.
With the existing models, we have already substantial use cases, and there are numerous unexplored improvements beyond the LLM, tailored specifically to the use case.