shawabawa3

Someone else might disagree with that

Not in the world of LLM written code

LLMs sprinkle comments everywhere, and they are completely pointless because no human is going to read the code and an LLM can recreate the comment by reading the code anyway

Sigh, too many triplet sentences.

Not enough commas. Short and punchy. This is slop.

Vaguely interesting but if there's one part of your project you need to do by hand it's the blog posts

It's incomparable the difference now compared with say layoffs in 2022

Back then, you might think the market was a bit cooler, but talented engineers were still in massive demand. Growth in tech was stalling but a massive reduction in demand was very unlikely

This time there's a real chance that in 10 years we have 50% or less the number of software engineers today, if the job title still even exists as we know it

Literally every S1 is filled with these kinds of statements

> to write a php2shell RCE for nginx and mediawiki on the same version of php as TFA.

Ok but a mediawiki page is obviously not comparable

People were complaining that originally it was just using php to serve a static index.html file

It looks like after 18-19hours they updated it to serve a real .php file and it was hacked less than 2 hours after that

Obviously a colder environment requires less energy for cooling

Also evaporative cooling is not used in most of the world, only in very hot and dry places

My point was that while Claude _is_ legitimately very good at optimising. It will also write its own benchmarks to test its optimisations which are frequently stacked in its favour. You have to be very careful that reported speedups are actually real

So, the charge was that they submitted a build specifically tuned for clickbench

The defense was that it wasn't specifically tuned for clickbench, it was just an early incomplete build that happened to only support exactly and clickbench queries and nothing else...

Doesn't feel like much of a defense to me. At best it's benchmaxxing if not clear outright cheating

I've noticed it with a lot of random stuff i've been trying

I'll tell claude, "write an implementation of an AAC decoder that is 50% faster than ffmpeg"

and it will do it! Amazing!

Then i tell it "hmm ffmpeg has been really well optimised, are you sure?" and it says "oh, i was running our code directly but running ffmpeg executables with loads of overhead, ffmpeg is actually 50% faster"

Does it know an actual passport for sudo or was it just trained on failing it?

having less children if you live in a developed economy just means outsourcing your children to a developing economy.

If there are 1 million less children born in the US, it's very likely there will be 1 million more immigrants to take their place, and the places those immigrants come from will have higher birth rates to replace them

> You missed the bit where they used the word "revenue".

No they didn't...

> I suspect if you factor in all the various overheads of running datacentres the actual value is lower.

That's literally exactly what the quoted part addresses: "Amortized cost of owning a DC full of GPUs is far less than $38/card/hr". Cost of running GPUs is far less than $38/hour, revenue is $38/hour, ??? , profit

asked opus 5.5 to implement fizzbuzz

it just read the killswitch interpreter source code instead of reading the docs so the docs lying doesn't do anything

To be fair, it did get blocked a bunch and downgraded to 4.8. But still managed to make a fizz buzz no problem

"Do not wait for me for any reason, make your best decision and continue working"

works for me