eh_why_not

The degree-9 polynomial, said to be a thousand times better than the original Taylor approximation in maximum error, also appears to be very close to the Taylor series in the first place.

Rounding the Taylor coefficients to 6 digits after the decimal:

1/3! = 0.166667

1/5! = 0.008333

1/7! = 0.000198

1/9! = 0.000027(56)

The first 2 are exact, the third is 5 digits only (so 0.000190), and the fourth is more different starting from the 6th digit (0.000026019).

The delta in the 9-th order is expected if you were to truncate the Taylor series starting from the 11th order to infinity (+ x^11 / 11! - x^13/13! ...).

> even though near the speed of light, more and more of that energy goes into higher mass and slower time?

This to me does not sound very different from saying the (relativistic) mass to be accelerated is larger, and is what's stopping a 1G (or any constant) acceleration to be maintained to bring the ship to the speed of light.

The concept may have been an overly simplified pedagogical tool though that I need to upgrade from. (I'll keep the comment chain intact since it's educational to me).

Question from someone with mostly back-end/services non-cloud dev skills: what exactly are Cloud Native skills?

For the comments saying we "just" need to maintain a 1G acceleration, should point out that as the ship approaches the speed of light, its mass increases [0].

And as the mass increases, so does the thrust required to maintain that acceleration. So that engine better be able to tap into some magical energy source, otherwise it can't maintain the acceleration at the higher speeds. :)

[0] https://en.wikipedia.org/wiki/Mass_in_special_relativity#Rel...

Edit: this interpretation may not be entirely accurate (comments below).

The redditization of HN: activism vs critical thinking.

The statement quoted above, however, is childish.

Hmm?

Question for people who prefer QR: how do you know you're not getting priced higher than other customers?

I wonder if the sources it lists are completely authoritative to its answer (i.e. the answer adds no new/hallucinated stuff that's not in the sources), or whether the sources are just "relevant".

Mind sharing how (tools/setup/etc)?

I have many shows purchased there and would like to have them on disc instead.

If there is lack of concern for the user and the long term, how can work be "well-intentioned"?

Intentioned for whom? And why should the public perceive it as good?

Who else had the commercial/totally-for-profit interest in slowing OpenAI down, and was in a position to do something about it?

It is hard to see what unfolded as anything but active sabotage.

YouTube is not the creator/owner of the content. There are even channels that are actively against the advertising business model and that rely on off-platform subscriptions and donations. I.e. the creator still gets value even if no ads show on their videos.

YouTube still shows ads on those channels though. And the more eyeballs and YT subscriptions these channels get, the more YT makes money off them.

Now what percentage of viewers use ad blockers? 5%? 10%?... YT still makes money off the rest. It's not like those are a worrying cost center to YT. Why then go to those extremes to shave the last penny of the last person who really strongly does not want to see the ads (which potentially undermines the advertiser's purpose of making you like their product)?

What are some good online services that can take a URL and give back a few bullet points? (which would then help decide whether to spend 20 - 30 mins to read the arguments)

Meta and Google are exiting "detailed ad targeting"? How so?

The Lick Observatory began operation in 1888 [0] (not 1988), and was constructed between 1876 and 1887 [1].

[0] https://www.lickobservatory.org/about/

[1] https://en.wikipedia.org/wiki/Lick_Observatory

It's crazy that the company's action is legal.

It's crazy how much power asymmetry exists between corporations and individuals nowadays. From forced arbitration (even for basic things like rent contracts) to non-competes, to now TRAs...

> But regulators have begun to take action on the legality of T.R.A.s. In the last year, the Biden administration has moved to limit the agreements.

Good to know. But solving problems at the executive level, without codifying the solution into law, makes it temporary and up for grabs by the next election, you know, for those jonesing for "less regulation".

The idea is that as long as it is coarse-grained geography it was less damaging.

But would that make it dangerous for political and ideological mass "influencing"? Maybe you're right. If we ever get a chance to have laws and regulate it, maybe allowing certain categories only.

The moral line is that you do not spy on people in order to "improve" their "ad experience". Geography is one way to do it that's then acceptable.

> Why should men see ads for tampons? Why should women see a message telling them to get a prostate cancer checkup?

If a man enters a lingerie shop, he will expect to see ads for tampons. If a woman enters a man's cigar club (or whatever), she will expect to see ads for a prostate cancer checkup.

"Why" is because the only possible way to do otherwise is by spying on them. Anything that requires customizing ads per person inherently requires the spying.

> There's all sorts of value and productivity that can be gained from targeting ads.

And there it is. You see it as an optimization problem of "value and productivity", and there is no place for privacy in that equation.

Much like a factory dumping its chemical waste in the town's river. A lot of "value" is created during the process.

No. The most ad targeting that should ever exist is by geography, and that's it.

Advertising has value in letting people know about your product. If you're selling locally, you rent a billboard on some street crossing, or a banner at a mall, or you buy a sponsorship at a local event because you think your target demographic will be there. At a demographic level, not individuals.

On the Internet, geography translates to maybe using zipcode from an IP address, and demographics translates to choosing which websites ("publishers") to put ads on.

And that should be it, that's fair and should get you enough bang for your buck.

Putting cameras in people's homes, listening to their conversations, following them around in their cars, stalking their relationships, tracking their voting records, etc should not be an acceptable way of increasing the efficiency of the targeting.

Of course you would make more money with this kind of spying (as what happens today). But this kind of invasion has great societal costs to privacy, and subsequently to freedoms, and should never be justifiable.