Hacker Newsnew | past | comments | ask | show | jobs | submit | nroets's commentslogin

Some changes to the algorithms and implementations will allow cheaper commodity hardware to be used.

There will always be an incentive to scale data centers. Better algorithms just mean more bang per gpu, not that “well, that’s enough now, we’ve done it”.

In 2023 I stayed in a capsule hotel in Bilbao. There I was told at check in that no talking or loudspeakers of any kind is allow in the capsules. That can only be done in the lounge area.

I spent 3 nights and got much better sleep than I would have at a hostel.


Could there be a motif unrelated to ICE ? That Home Depot does not like that day labourers are loitering and approaching customers entering and leaving the store.


I believe Home Depot offers a similar service now so in a way they are directly competing


Likely because they contrast with many of its own employees' lack of helpfulness, knowledge, or work ethic.


if so, you wouldn't expect this to be a new policy


What studies ? What data ? David Bessis basically says that there are so few twins reared apart that scientists can't make definitive conclusions.


I don't understand why you are challenging me here?

Isn't your question exactly that addressed by the (admittedly too long) article? That the graph Paul Graham presented proving the dominance of inheritance wasn't based on any science or data?


Your comment mentioned "studies" plural.

There are many studies of twins that try to determine if genes influence intelligence.

Some look at twins who are raised together. One [1] concludes that "MZ (identical) twins differ on average by 6 IQ points, while DZ (fraternal) twins differ on average by 10 IQ points".

[1]: https://pmc.ncbi.nlm.nih.gov/articles/PMC6202166/


> Your comment mentioned "studies" plural.

Yes? I mentioned them because the article was about a bunch of studies? I was asking the poster why you would not be interested in the validity of such studies and just decide that "common sense" was enough to make a decision?

The question was asked with genuine curiosity as this forum is mostly filled with people who appreciate science and empiricism. And I was hoping there could be a reasonable discussion.

But I'm out. An interesting discussion should be possible here purely based on data and statistics but clearly - from the downvotes - that I've stepped into some toxic American identity politic minefield.

I learned quite a time ago that it's risky to raise certain scientific subjects with USAians including my US relatives: biological evolution, the science of climate change, renewable energy or justifications for gun control - without the conversation getting emotional and heated. But I still find it weird.


The article only asks the question of scientists have data to conclude that IQ is inherited. The author is only saying that there are so many problems with the little data we have, that he cannot rule out correlation without causation.


So you point to one instance of highly targeted sabotage aka sanctions. But Snowden and others exposed many instances of espionage dragnets.


You can buy (go long) a BTC future with only $10,000 or less of collateral. So you can get lots of leverage.

Another reason is that the future may be trading slightly below the spot price of BTC due to lots of traders shorting.


Here's a fairly fast algorithm: Look for a number that appears in both { j⁵+k⁵+l⁵ } and { i⁵-n⁵ } where 0 < j <= k <= l and 0 < n < i

If we only consider l and i under 250, then the sets would contain less than 3 million integers each.

Strength reduction can be used to replace all the multiplications with additions: https://en.wikipedia.org/wiki/Strength_reduction


It's likely that the original search also used strength reduction to save a lot of cycles (effectively replacing all multiplications with additions):

https://en.wikipedia.org/wiki/Strength_reduction


"Theft" means taking something from someone without consent. Who lost what ? There is no law suite, so maybe it's a donation ??


The taxpayers / government. If they have been abusing their NP status to avoid taxes, they should have to back pay those.


What taxes are they not paying?

I am unable to find any concrete claim of specific tax avoidance. Only these exasperated “but taxes” comments.


Non-profits are literally tax-exempt. OAI spent 10 years being tax-exempt in exchange for doing work that fully benefits the public. Now that work, 10 years of tax exempt work, is being handed over to a taxable outfit, a for-profit organization. If the result of 10 years of tax-exempt efforts get handed to a for-profit company, the taxes that were never paid should be because the public benefit that got them the tax benefit wasn't fulfilled, in fact it was stolen and handed to ultra-wealthy capitalists.


You mean the results that a few other companies almost instantly copied and productive themselves once the way to do it was discovered? There is no moat around LLMs.


What taxes did the non-profit skirt?

All the sources I can find say that the revenue of ChatGPT was through the for-profit division, and that they’ve been paying taxes on all their revenue.

Is there some other tax that they’ve avoided paying?


Sales & property, see my nearby comment for links


Oh shit, the company that revolutionized AI didn’t pay their fair share of SF property taxes. Now I understand the outrage!


It doesnt matter what kind of tax they didnt pay. They SHOULD pay tax. Otherwise this makes it a loophole for private companies to dump research & development costs on the taxpayer but reap all the profits.


Huh? The loophole is already there.

Everything of their restructuring was signed off on by multiple states’ attorneys general. And their for-profit entity pays taxes like any other company.

Making them pay tax on stuff they did while a non-profit is making up laws on the fly - a strong, rule-of-law-based system is critical for the US to function properly.

You can’t just arbitrarily make decisions based on what you think should happen because it’s fair or unfair.

If you want OpenAI to pay back taxes, you need to change the laws first.


The issue is not the laws. The issue is that OpenAI mislead officials and externalized costs on the taxpayer.The extent to which this happened should be looked into by professionals.

It's not about changing the laws, it's about enforcing the ones we have fairly. Too many orgs and companies buy politicians, and now ballrooms for them


What law was not enforced?


People who are experts should look into it and let us know (i.e. an investigation)

I would not trust the corporations and politicians to be forthcoming or transparent on this


Do you think really think they were profitable during that time?


Surely they have never turned a profit and are a long way from being profitable. If so, what taxes, current or back, would they owe?


Income taxes are not the only tax non-profits are exempted from. Sales and property taxes are others, depending on jurisdiction, California being one such state. I am not familiar if OpenAI-NP has been exempted from these

https://www.fplglaw.com/insights/california-nonprofit-law-es...


The money they received was tax deductible for the people who “donated it”. They money should have been taxed as income for either the earner or OpenAI.


I'd be curious if people were actually writing off their OpenAI bills as donations. That would be a big number for the enterprise deals, if they qualify as a donation


Surely the money coming in would otherwise have been investments exchanged for stock, which are not taxed until gains are realized.


The AI researchers who joined and worked for less money than they would have been paid by a big tech company because they thought it was the right thing to do.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: