gblargg

8 ms·

151.
by gblargg 4mo ago
Sounds like this one just silently corrupts the results. It's more like when YouTube shadow deletes your comments without any notification, it's just gone after a few minutes.
152.
by gblargg 4mo ago
Seems like this will backfire. Now when developers encounter problems with Claude Fable, they will have an easy explanation: it did it deliberately and intentionally vaguely. There's no way to falsify it. It's reasonable to expect
153.
by gblargg 4mo ago
Reminded me of this fun video of a guy spinning up a water turbine generator and getting it synchronized with to the grid the old-style needle phase meter, then connecting it successfully: https://www.youtube.com/watch?v=xGQ
154.
by gblargg 4mo ago
The ELI5 version is you're arm-wrestling someone and they suddenly let their arm go limp, so your arm slams down on the table since you can't react that fast.
155.
by gblargg 4mo ago
Charge them the cost of installing mitigations that can pick up the load if they drop it suddenly.
156.
by gblargg 4mo ago
Maybe he needs to optimize the frying part now. Looks burned to me.
157.
by gblargg 4mo ago
https://archive.is/c8Pep (avoid the popups)
158.
by gblargg 4mo ago
I educated anyone who asked about the NTSC filter over the years because I wanted to see less-optimized implementations of it, given how much faster hardware is than the mid 2000s (it precalculated the kernels for every color at every phase
159.
by gblargg 4mo ago
Modern processors are pipelined, where they run a lot in advance of when the result is needed. A mispredicted branch requires throwing out all the advance calculations on the incorrectly followed path. The branch predictor can't predic
160.
by gblargg 4mo ago
It only increments if the number was less than 500, effectively just saving the ones less than 500.
161.
by gblargg 4mo ago
It would be funny if someone hacked the database to block any praise of the current politicians and this censorship, only allowing criticism of it.
162.
by gblargg 5mo ago
As an aside, using an LLM results in a post that's all black text, except a couple of colorful emojis that jump out: > Speak with confidence. “I think that I can possibly consider agreeing” . Don’t share weak decisions or opinions.
163.
by gblargg 5mo ago
I'm surprised their 1GB file wasn't cached entirely in RAM during the attack, eliminating the SSD from any timing. Do people keep their machines that heavily loaded that a file being constantly read from doesn't stay in the c
164.
by gblargg 5mo ago
Or mine where the entire browser profile is on a ramdisk.
165.
by gblargg 5mo ago
If you compressed it with H.264, it wouldn't make much sense to send it remotely to be encoded with a better codec.
166.
by gblargg 5mo ago
The context was remotely encoding the video, which would require sending the uncompressed stream.
167.
by gblargg 5mo ago
I'm reading that L1 takes 4-5 cycles to read on modern CPUs, whereas it was just one cycle in the late 1980s.
168.
by gblargg 5mo ago
It kind of reminds me of the equality of opportunity people versus the equality of outcome people. One sets the starting conditions for developers, the other the ending conditions for users.
169.
by gblargg 5mo ago
BTW there's a summary today that has bullet-point markup and it just wraps around without proper formatting. Title is "Show HN: Tiny-vLLM – high performance LLM inference engine in C++ and CUDA".
170.
by gblargg 5mo ago
I declined things like rm -rf because the path was relative and it wasn't showing me the current directory. How would I know what project it was in?
171.
by gblargg 5mo ago
The individual franchises can always get the clue and drop being a franchise, just be a local store.
172.
by gblargg 5mo ago
I tried the other person's videos and had to stop after about five seconds. Mr. French's was a good watch.
173.
by gblargg 5mo ago
*grey, since it was Star Wars
174.
by gblargg 5mo ago
Unfortunately this makes every Bricks and Minifigs store a risk to do business with, selling or buying. You're either supporting a criminal enterprise or risking being a victim of one.
175.
by gblargg 5mo ago
AI is downgrading non-AI search results with AI slop pages that get ranked high.
176.
by gblargg 5mo ago
Years back Google temporarily put their old database from the good times up for searching. You could actually find technical results reliably. It was so nice.
177.
by gblargg 5mo ago
They need to have a way to report AI videos not labeled as such, AND a checkbox to filter out AI videos on the home page and in search results. Not holding my breath for either.
178.
by gblargg 5mo ago
You're missing the lack of verification. Normal computer tools can be verified to work reliably. Cat can be tested to copy data. Sort can be confirmed to sort properly. You can't verify that a coding agent can reliably produce cod
179.
by gblargg 5mo ago
Perhaps it's good to know which one, like knowing what harmful ingredients are in your food so you can avoid ones you know are unsafe.
180.
by gblargg 5mo ago
It shouldn't be crossing solid white lines as a matter of routine.

More ›