Typically maintenance, reliability, support and liability; nothing new in SaaS land. Lots of companies have the money to hand-roll things that are provided by some external SaaS but would much rather it was someone else’s problem.
Also for creatives, some sort of aesthetic control? Most of the examples on this page look like templates you’d see in an office suite that appear swish, but are ultimately bland.
This one is pretty terrible (right after “The worst outcome is a spurious rerun, never lost work.”). We’ve got lands, several "no X", hyphenation, strange noun/verb sentence order and an unnecessary analogy word (swallowed).
> There's no read-modify-write and no truncation of the dirty period, so a write that lands during a job can't be swallowed.
That's the main reason I'm using GPT models. I'll ask Fable to analyze something, then pipe its output straight through Astra without even looking at it first.
A fun demo of this from earlier this year (Tsoding): https://www.youtube.com/watch?v=qjWkNZ0SXfo The good stuff starts around 7 mins, but it’s a great presentation and it’s almost magical how everything comes together.
Personally I think the price is way too high right now. It’s a power hungry gaming GPU. The efficient single card equivalent would be a 4500 Blackwell which launched at about $3500. Or you could get a 9700 32GB or an Arc B70 for well under $2k, today. You only buy a 5090 if you want absolute speed.
32GB is still not that much. I would rather get a Spark and have the RAM to experiment with larger LLMs, even if it was slow.
How does this relate to 4500 vs 5090? I'm just pointing out that 5090 likely has twice the performance of the 4500 and likely maintains that at 2x watts if you want.
You didn't specify in your earlier post, so I wasn't sure exactly which comparison you were making. But yeah, the perf/watt actually looks the same for those, so the cost per token evens out. It is nice not having to manage 400+W though. I like the 4000 for that reason, it's effectively a 3090 that runs at half the TDP.
I don’t think there is any consistency in behavior. People react viscerally to the unproven belief that the Facebook app records conversations. They’ll also happily buy big TVs despite the fact they’re very much listening to you and we have solid evidence that it’s for ads or residential proxying. Most people have no idea how ads “follow” you, or how companies can figure out what you talked about by connecting the dots from other user metadata.
It’s time and commitment for sure. On the mental side - 30s is still young. Most US academics and doctors are barely out of grad school at 30 and their careers (which involve continued learning) are just starting.
Approach it differently as a salaried adult. You can pay for better tuition on your own schedule.
But you can find time. I listen to Michel Thomas or other courses while I’m walking to work. Put on headphones and ignore other people, they aren’t listening to you anyway. Other things like flash cards you can do any time you’ve got 10-15 mins.
> Technical documentation, where what the reader wants to come away with is the clear facts on what something does and how to use it/whatever, is a different matter.
I don’t agree with this one. Writing technical documentation can be a slog, but quite often you will discover very annoying design decisions in the process. Like you’ll write a tutorial for a CLI and realize that actually some flag makes no sense, or some default should be changed.
This also goes for dogfooding. If you’re using AI to write and run the tools, try to use them yourself and see how nice they are. AI will happily run whatever jank it’s working with.
Docstrings perhaps, but even then I find myself editing heavily to avoid cruft sneaking in.
I like Elicia White’s Making Embedded Systems because it covers a lot of practical stuff like how to read a datasheet. It’s a good book for existing programmers.
A lot of electronic books struggle to find balance between super easy projects, very specific cookbook recipes or a laundry list of components.
https://wiki.factorio.com/Blueprint_string_format
(Also unrelated, I think by clearance they mean tolerance?)
reply