HansHamster

No, that was directly from deepseek. And it happened very consistently (3 or 4 times in a row, clean session every time, and across 2 harnesses). Guess I'll give it another try when v4.1 flash is available.

I must be doing something wrong. I gave v4 pro a try a couple of days ago, gave it a simple prompt like "clean up functions x and y in file z" and it would always start off promising, just to quickly get sidetracked, start hallucinating problems in the code, and just get stuck for hours until I interrupt it:

— hmm — 0x2D696370 — little-endian bytes: 70 63 69 2D = 'p','c','i','-' — hmm — WAIT — WAIT — !!!!! — *WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — *HOLD ON — HOLD ON — HOLD ON — WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — !!!!!!!! — *WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — *OK — WAIT — I THINK I FINALLY SEE THE WHOLE PICTURE — I NEVER READ IT — AND — THE LAYOUT — hmm — !!!!! — *WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — WAIT — HOLD ON — HOLD ON — HOLD ON — HOLD ON

Then gave the same to Sonnet 5 and it was done 15 - 30 minutes later. I tried v4 pro both in claude code and codewhale with similar results. Haven't tried the new deepseek harness.

Interesting. You can see the building from the beach and I always wondered what's inside. And while it's the only (decommissioned) nuclear power plant on Long Island, it's not the only nuclear reactor. There was also the High Flux Beam Reactor at BNL that was decommissioned in the 90s:

https://www.bnl.gov/hfbr/

https://www.bnl.gov/hfbr/hfbr-complex.php

Yeah, I'm still concerned about crystalline entities suddenly showing up. Have they ever fixed it? I don't see anything in the issue tracker, probably because no one was left last time to report it...

Maybe their CRTs had horrible burn-in and they had to erase everything 9 times before it was gone...

I would expect all RNG algorithms to be deterministic and stable with their seed, but the cryptographically secure ones to have some additional properties like making it unfeasible to reverse the seed from the output, having a very long period or strong guarantees on the distribution of the output. It's just that using a 'secure' algorithm is often overkill for a game when you don't really need those extra guarantees.

It feels a bit unfair to say that it is faster by being able to tell the total length from the first byte and capping it at 64 bit, while some of the other formats can store arbitrarily large integers. I guess you could use another variable length encoding for the prefix at the cost of some performance and using even more space...

Eh... unfortunately they shut down their forums a couple of years ago. So good luck getting any form of support as a free user even if you run into real bugs in their software (believe me, I tried...)

That being said, I have used their ice40 and ECP5 FPGAs with Yosys for a couple of small projects and that worked perfectly fine.

And Hamamatsu (and some others) still produce and sell photomultiplier tubes. The microchannel plate PMTs are pretty nifty things [1]. You can get single-digit picosecond time resolution out of them. [1]: https://www.hamamatsu.com/us/en/product/optical-sensors/pmt/...

bacon!

Clearly, we should place a radioactive source, a Geiger counter, and a computer in a sealed box. Every time the counter registers a decay, the computer performs a HTTP GET. Thus, we end up with Schrödinger's webserver...

At least you can tick the "stay signed in" checkbox and... get kicked out a few hours later with a smug "you successfully signed out" message.

Obligatory Futurama reference: https://www.youtube.com/watch?v=N4vIBijzg4w

> which is faithfully translated by the (hopefully bug-free) compiler.

"Hey Claude, translate this piece of PHP code into Power10 assembly!"

> You have to be the one who reaches out.

But that's the whole issue. Who am I supposed to reach out to? The 2 people at work I occasionally talk to because they happen to sit in the same office as me?

Funny how it went from "just get an Nvidia card for Linux" and "oh my god, what did I do to deserve fglrx?" to "just get an AMD card" and "it's Nvidia, what did you expect?"

> Recently, GPT informed me that the strong force is really a tiny after-effect of the "QCD force"

Maybe you should not take everything GPT tells you at face value? I have no idea what this QCD force is supposed to be. The strong force is _the_ force of QCD. The Standard Model still considers the electromagnetic, weak and strong force. The description of the weak and EM force can be unified into the electroweak force and there are theories that try to also unify it with the strong force and even gravity, but there are issues on the theory side and no clear evidence on the experimental side as to which direction is the correct one.

The Standard Model and General Relativity are still our most successful theories. It is clear that they don't tell the whole picture, but (annoyingly?) it is not clear at all where this is going.

Just for dark matter there are probably a dozen proposed hypothetical particles, but so far we have found none. But maybe it's something completely different...

that was the joke

Should be fine as long as the chips have ECC

Isn't Incus/LXD separate from and running on top of LXC? People sometimes seem to use the names interchangeably which can be annoying because I run just plain LXC but when looking stuff up and come across "this is how you do XYZ on LXC" they are actually talking about LXD and it doesn't really apply. I can't recall what is was last time, but this has happened a couple of times already...

Meanwhile, the DPO4054 I use at work has issues with half of the buttons, the data returned by some commands doesn't match the manual, and the probes from some newer scopes don't fit even though they all just use BNC? Maybe the software on it is just outdated? Nope, the only version you can get from the Tektronix page is older than what is on the scope. And there are newer versions, but they are only for the DPO4054B and not the older DPO4054...

This also confused me. The current ones have very distinct colors and also all the previous series used different colors as far as I can tell: https://en.wikipedia.org/wiki/Banknotes_of_the_Swiss_franc

This brings back memories of my old HP laptop with an Athlon 64 and a Radeon X200M. The crappy FGLRX driver only supported overlays (afair) and so when running something like Compiz it would transform the window with the green background but the video itself would stay in place and it would just stick parts of the video on top where it happened to overlap. I still remember being excited when the open source drivers finally gained support for r300 and could do proper textured video...

Can confirm. My 3TB Seagate was the only disk so far (knocking on wood) that died in a way that lost some data. Still managed to make a copy with dd_rescue, but there was a big region that just returned read errors and I ended up with a bunch of corrupt files. Thankfully nothing too important...

> When you send your design to pcbway or jlcpcb they have much tighter control over the process, so you no longer have to worry about this stuff.

Funny that you mention jlcpcb. The last time I submitted a board with tight differential pairs (but still within their listed specs) to them they basically told me to increase the amount of copper, so I assume they had some quality issues in the past:

> we have new rule since Dec, 2022, if the copper areas are less than 30% of the board in each copper layer, the space between trace and trace should be at least 0.15mm to avoid short circuit.

So I had to add a few copper pours and everything was fine :)

> Including things you like quite a bit, such as WiFi support A while ago I set up Win 98 on an old P2 laptop from 1998 I found in the trash at work. Even that can do WiFi with WPA2. It was a pain to set up and requires 3rd party software to establish the connection, but aside from that it works just like any other network card.

You mean Microsoft Windows which dropped support for Zen 1 with Win11 not even 5 years after Zen 1 was released? Meanwhile, Linux will still run on a 30+ year old CPU...

This is truly amazing, fractal levels of nonsense. I can't stop reading and I'm waiting for the "hah! gotcha!", but it just goes on and on...

It's t̶u̶r̶t̶l̶e̶s̶ wasps all the way down!

> A sleeper train takes 0 useful hours, because you’re asleep.

Only if you manage to fall/stay asleep...