this post was submitted on 23 Jul 2026
25 points (82.1% liked)
Showerthoughts
43021 readers
146 users here now
A "Showerthought" is a simple term used to describe the thoughts that pop into your head while you're doing everyday things like taking a shower, driving, or just daydreaming. The most popular seem to be lighthearted clever little truths, hidden in daily life.
Here are some examples to inspire your own showerthoughts:
- Both “200” and “160” are 2 minutes in microwave math
- When you’re a kid, you don’t realize you’re also watching your mom and dad grow up.
- More dreams have been destroyed by alarm clocks than anything else
Rules
- All posts must be showerthoughts
- The entire showerthought must be in the title
- No politics
- If your topic is in a grey area, please phrase it to emphasize the fascinating aspects, not the dramatic aspects. You can do this by avoiding overly politicized terms such as "capitalism" and "communism". If you must make comparisons, you can say something is different without saying something is better/worse.
- A good place for politics is c/politicaldiscussion
- Posts must be original/unique
- Adhere to Lemmy's Code of Conduct and the TOS
If you made it this far, showerthoughts is accepting new mods. This community is generally tame so its not a lot of work, but having a few more mods would help reports get addressed a little sooner.
Whats it like to be a mod? Reports just show up as messages in your Lemmy inbox, and if a different mod has already addressed the report, the message goes away and you never worry about it.
founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
This vastly underestimates how big and interconnected modern LLM networks are. They are 100s of GBs in size, and transmit GBs of data between layers 100s of times per second. Having to transmit any part of the intermediate state of an LLM over a 1 Gb/s link would degrade the performance of an LLM by over 1000x. Current top-of-the-line datacenter GPUs are running about 4 TB/s total memory bandwidth, a limit LLMs are already pushing up against.
The kind of clustering you're talking about is being done: between multiple GPUs on the same machine. The slowdown of dropping to the measly 64 Gb/s speeds of the PCIe 5.0x16 bus is a huge performance hit and has to be done very carefully. The various parts of the computation are broken up and shuttled around.
It's the combination of massively-parallel compute (in the GPU cores) and insanely-fast, fully interconnected RAM that makes generative AI possible. The load characteristics aren't conducive to parallelization.
Not on tiny nodes separated by tiny fiber pipes... as I said elsewhere: a $100K node sucking down $1500/month in electricity could start to serve some useful loads, and provide all the hot water you could ever need.