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

Most have their pension automatically track index. Nasdaq changes rules so that the pension funds are forced to buy stock while very volatile and likely overpriced.

I'm willing to bet that most funds will just change their reconstitution process to give themselves a much longer period to add new IPO stocks to their portfolios and end up avoiding most of the drama.

On the other hand if they don't I'm making popcorn because the lawsuits and political fallout if / when this goes wrong is going to be epic.


Those poor pension funds, unable to make their own decisions. Forced to keep all of their money in $NDX forever.

It must be terrible to be so utterly powerless.


That's not how pension funds work though. Pension funds are extremely conservative. They invest on horizons spanning decades. They can't just make decisions from one month to the next. The very short time frame of an index accepting Tesla is the problem. It's like telling a train "quick, make a right here"

And how do you suppose these rules were originally introduced? Do you believe that the reasoning was particularly solid then, and is it directly applicable in the current situation?

Yes. These rules were introduced, because that's the purpose of pension funds. To invest in very long-term ideas and assets. That's their whole purpose. I'm not sure if we're debating here about the purpose of pension funds, or their strategy of investing in index funds. The first one doesn't change, the second one, maybe.

Pension funds use (you guessed it) index funds like your average passive investor.

Sure we can hate pension funds but fuck the management of those indexes.


Pension funds are perfectly capable of switching to a different product if they don't want index funds that reflect the market. But they do want index funds that actually reflect the market, so they want these rule changes.

The management of those indices is just doing exactly what the vast majority of their customers want.


What? Absolutely not. They are doing what the President is telling them to do.

The customers of indexes are providers of funds (Vanguard, BlackRock, ...). In turn, people like you and me are clients of those providers.

Indexes are changing the rules thanks to lobbies by the world's richest people so they can get even richer by dumping their overvalued stocks.

If you are not a SpaceX investor, you are losing money. Including pension funds.


>What? Absolutely not. They are doing what the President is telling them to do.

This is a paranoid delusion.

These rules were never on particularly solid ground, it's only natural to see them gone when you have significant upcoming IPOs changing the structure of the market. It is the job of these indices to keep up with that.


I'm sorry but you're very naive. The rules were changed because it was advantageous to SpaceX and the richest man on Earth.

Those rules were implemented to protect the index from being skewed when an IPO happens because the IPO is exactly the moment where the stock may be overvalued and very volatile (compared to all the other stocks that are in the index). We especially don't know at IPO time if the company is changing the structure of the market because it's too soon. And in fact, we won't know for a couple more weeks or even years; but by that time the price will have stabilized.

The index is supposed to be slow at rebalancing to minimize costs and make it practical for providers as well.

And as I write this comment, I've just read that they cancelled the changes[1]. Which completely proves your last sentence wrong.

[1]: https://press.spglobal.com/2026-06-04-S-P-Dow-Jones-Indices-...


>I'm sorry but you're very naive. The rules were changed because it was advantageous to SpaceX and the richest man on Earth.

The rules were changed in the mid-90s because of a significant amount of big IPOs which did not have a reasonable path to profitability. That's not even close to the situation we're in right now. Both OpenAI and Anthropic are clearly solid businesses.

> We especially don't know at IPO time if the company is changing the structure of the market because it's too soon. And in fact, we won't know for a couple more weeks or even years; but by that time the price will have stabilized.

This is largely magical thinking. How many more weeks should I wait before $TSLA will stabilize?

>And as I write this comment, I've just read that they cancelled the changes[1]. Which completely proves your last sentence wrong.

Or maybe deranged Americans just turned this non-political issue into a political one, and S&P global thought this was not a hill worth literally getting shot in their office for? I don't know, that'd certainly be my primary concern in their place. Wouldn't it be yours?


> The rules were changed in the mid-90s because of a significant amount of big IPOs which did not have a reasonable path to profitability.

Could you give me a link to read up on that?

> Both OpenAI and Anthropic are clearly solid businesses

Maybe you're an insider so you have more information than me; I'm not so I don't know.

> How many more weeks should I wait before $TSLA will stabilize?

TSLA is mature enough since their IPO to be included in an index.

> Or maybe deranged Americans just turned this non-political issue into a political one, and S&P global thought this was not a hill worth literally getting shot in their office for?

No, it's probably because major fund providers and other market participants (including retail) saw through the subterfuge. Indexes and providers and retail investors have nothing to gain from the former proposed changes. It was very obviously a play by SpaceX to force passive investors to buy into the stock, therefore pumping the price up.


> There’s no skill file that contains the tacit knowledge of a person who has reconciled a thousand payrolls.

That person has zero skill in actually making tight automation that doesn't just fall over. And I have yet to see an AI agent that tells them "look, your requirements are contradictory, given this and that, these two cannot coexist".

Those little sycophants will just go and try to please the domain expert and placate him in all ways possible. Bend backwards rather then forcing them to reassess their assumptions.


This is a terrible benchmark. It literally tests the models on their ability to track shifting line numbers. If they cannot keep up, no amount of abstract reasoning can redeem them.

Where did you get that idea? It uses mini-swe-agent, same as SWE-Bench.

https://github.com/datacurve-ai/deep-swe


Foraging was so good we only had to work like 2h a day. We had to start farming simply because the population growth from foraging was exponential and we have ran out of easy-to-forage ecosystems.

We've also ran out of freedom at that point. We used to simply walk away from bullies, but that was no longer possible, with the farm and all.

At some point bullies stopped charging for racket. They stole the land and told us we are now working under them, for them.


We always hold one at $WORK at the end of the year. I am not a fan of racing nor CS (prefer UT) so I went and made a retrogaming corner last year.

The year before that the Among Us was the highlight. I landed on imposter so many times they started ejecting me habitually. :-D

I think this year I'll finally play some AoE2 with my wife. We're going to Wololo the shit out of my colleagues. Well, mostly her. I'll just send her some tribute I guess.


Not OP, but I routinely load 150k tokens into context. A full sub-package to work on, select other files in the monorepo, e.g. front-end visualization and back-end data loader. Then work some 150k tokens, then start again.

At the end, cache hit rate is like 99.5% if Novita is not having issues.

For official DeepSeek API, 99.9% or something.

Custom harness that never compacts or otherwise doctors the history.


Those numbers make sense to me...120 million input tokens is like 120 sessions of hitting the full context limit, which seems like a lot to me though


It's a vasoconstrictor. Gets more blood up to the brain.


That's called LUKS2 and it's the default on Linux. You just type passphrase on boot. It's not tied to the motherboard.


What if you forget the passphrase after not using it for many years and you suddenly need a file on the drive?


Print it on a piece of paper and put it in a lock box.


Better still: LUKS allows you to set up multiple entry keys, so use two, either of which will grant access to the drive.

* Your preferred memorized passphrase and will never be written down anywhere.

* A random key you can print and store in a box somewhere.

Then if your backup paper gets lost, you can revoke/replace it without having to abandoned your memorized favorite.


Yep. You can also put your key on a usb drive that can be read on boot.

Just choose a good quality one....


A few ideas for extra security:

* Split the recovery key in two, store each half with a different friend. (If you're feeling fancy, XOR the halves and store that with a third friend, then any two out of three will work.)

* Sneak the key into something you know friends/family won't throw away while you're still alive, like stuck to the back of a sentimental photo in a frame.

____

That said, I think I'm wandering from the original "accumulating dusty old drives in a box" scenario, which has a simpler solution: Keep a growing old_drives_keys.txt file on your current (encrypted) main device.


Yep, this is the way. It survives human memory and doesn't depend on software.

If you keep it in a dark environment that's not super humid the ink should last a really long time. Even in non-optimal conditions (NY summers with high humidity, etc.) I've had regular pen ink last for decades with no signs of fading away.


Except weights will be unstable - temperature and frequency dependent - and we still have issues delivering analog circuits reliably to the spec. So it would take multiple attempts.

But yeah, as soon as the digital models start to plateau, ASICs and then this will happen.


You could SFT for temperature and frequency stability.

I'm not even kidding. Modern ML systems already eat errors - what's one more error type for them to eat?


I am 100% certain it's a sigmoid. Speed of light is finite. :-)


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

Search: