I think over the decades I got one RAM disk fail. And another one didn't fail but, although I bought a pack of four sticks (2x x2), was detected (by memtest, when building the rig), as having another model name than the three others: even the shop who sold me the RAM was confused (don't know how that happened).

The usual component that I've seen fail the most would be the PSU? (before I started buying quality ones). One NVMe drive (an "ADATA") died on me even though it was nearly new.

PC assembled with quality parts usually last a very long time.

Ironically enough it's because of its bet on RAM companies that all of Situational Awareness' publicly trading stocks got bought by the highest bidder (Citadel) when the (leveraged) shit hit the proverbial fan and Situational Awareness created a hole of tens of billions.

Situational Awareness was not forced to sell its share in Anthropic for, by contract, investors in Situational Awareness have no right to force the sale (before this month of September I think) of shares in companies not publicly trading yet.

Citadel bought those RAM stocks right at the end of July, during the dip. Amazing timing for for example MU bounced back from $746 back to above $1000 (now around $950).

Now Situational Awareness may still do exceptionally well: they've got $5 billion or so left in Anthropic AIUI.

That is incredible. My front ones are toast after only 15 000 miles / 25 000 km ( Continental Sport Contact 3 / N0 rating) and two years. 95 000 miles and I'd need to change them not because of wear but because they'd be legally too old.

At some point I'll buy a reasonable car.

Same thing here brother: DCP-L2510D (laser printer + scanner).

Wife is on MacOS and cannot print/scan from there, so I'm the scanman.

But of course it's hard not to both marvel and lament (at) the fact that in the mid-1990s (so yup 30 years ago) I could:

    nc 192.168.1.150 9100 < tiger.ps

I know, I know, about garbled print jobs, pages of codes being printed, etc.

But still: it's hard to not at least wonder if some things haven't been lost. For a start printers with their own IP addresses were sweet (instead of a USB cable and, now, people setting up, say, a Pi just to expose a LAN IP for/to their printer). Then of course another talk could be had about the "quality" of modern day HP laser printers compared to what you'd get int the nineties.

> Later, the apps that would clone a USB dongle came along, and one software company that used USB dongles actually advised using this tool when we started running their software on Blade servers that did not support adding 24 USB devices.

That's a good one!

Many of us noticed that, on various topics: it happens once in a while. When there's sufficient interest again, things can bump and it looks like it's brand new but it's actually some kind of holy thread necro'ing.

And when it happens, there's also usually one or two people unfamiliar with TFA necromancing that posts a message not unlike yours.

Same

> Have you ever looked back on your history?

All the time? I search in my history, re-launch past commands, etc.

Now it's typically recent history, usually same day.

I know about the shortcuts etc. but I also need to at times pipe history into other commands so I've got a two letters alias:

    h?

And then, when the AMOC shall start running amok, we'll be forever in extreme cold.

Where should I flee?

I've got another car that's 13 years old and has got 126 000 miles / 202 000 km and that is 13 years old. In the 8 years I've had this one, my "green" doctor buddy changed four times his car.

Who's consuming the most? My ecological doctor friend who changes car every two years, my 13 years old daily or my 38 years car? (most cars have long been destroyed after 38 years)

Sure, the 38 old Porsche is a bit of a gas guzzler but the thing still runs strong. One Bosch Motronic electronic board for the fuel injection and that's it: no OTA updates, fully mechanical dashboard (which looks gorgeous btw).

Who's really the consumerist here? Me or my doctor buddy who buys a new EV every two years?

And if instead of an old Porsche I was using an old Toyota with much better mpg, then the equation is even better.

I'm not against switching to an EV but I don't want any of this OTA updates and 24/7 connected bullshit: give me an EV that doesn't phone home and I may listen.

Things however don't seem to be heading that way so I think I'll be driving used cars for a very long time.

I giggled.

So it's 10x. And no need for Unicode codepoints.

They've been acquired by AMD. Those saying the model sucks are completely missing the point: it was a proof-of-concept.

The question is: what happens to a model like Anthropic's Fable 5 that does, what, 70 tokens/s (and requires lots of output tokens) when the latest open-weights model is etched on silicon and does 14 000 tokens/s?

Shall the better model still have the upper hand or will the raw speed compensate?

I see what you describe all the time, because I do review the code the models do produce.

It's not just incredibly verbose: it's constantly missing that there's an obvious, elegant, small, way to solve what was asked and instead it goes ballistic and creates nonsense.

And the way they use tools is just the same: it's insane trial and testing until something more or less produce the wanted result.

I've explained it here already but the craziest I had was, like you, a one line test that was basically the following:

    if ( a >= 0xab000000 && a <= 0xabffffff)

But the model decide to go nuts: it noticed a pattern (just like it notices a pattern in your example) and decided to convert the native integers to strings to then do substring matching on the hexadecimal representation of the number.

I.

Shit.

You.

Not.

And all the people here who are saying that "it works" have no idea as to the amount of technical debt they're creating.

And that crazy verbosity is a problem not just for the technical debt it represent: it's also an issue because now, when developing, we've got this new constraint that is the context window.

It's a nice tool but it should be used with caution.

Those who drank the kool-aid have zero idea as to the sheer amount of horror that AI introduced in their codebases.

So higher-end Claude models (I've got a subscription btw) do work fine right?

What makes you think that in six months he cannot swap DS4 Flash 0731 for another model that could be equivalent to todays' top OpenAI/Anthropic models?

Or are you going to explain in six months that, after all, the top models from Anthropic from today are unfit for use?

I doubt it: PLA begins to soften at as low as 55 degrees Celsius. If there's weight on it, it's worse: the piece shall quickly deform and become unfit for its purpose.

This rig looks like it means business: I think PLA would fail.

If, like me, you cannot print ASA (say because you've got a printer that is not closed and that won't heat enough), then PETG-CF (PETG reinforced with some carbon fiber: it's got better heat deflection than plain PETG) is a safer bet than PLA for parts to put in PCs/servers/rack. Moreover PETG-CF do look really good.

And yet the entire thread here is people bitching that it's neither 5.6-sol nor Opus or Fable 5.

BTW why are OpenAI and Anthropic even releasing models like terra/luna and Sonnet?

Why? Just why?

Is there a... market?

For you can't have it both ways: either Sonnet and terra/luna make zero sense for Anthropic and OpenAI or Google is a player.

How magnanimous: they make the effort to support the OS their entire company is based on? It's not just for training the model and hosting them: it's also to run them, where they constantly use Linux to do their thing.

Without Linux, there's no OpenAI.

So I take it the Linux community should be so grateful that OpenAI supports Linux, right?