Justin Buist
@justinbuist.bsky.social
540 followers 1K following 5K posts
Software developer, data engineer, FIRST mentor, space nerd and occasional college student. I'm more than that but that's what I show online.
Posts Media Videos Starter Packs
justinbuist.bsky.social
Watts are a measure of power.

HTH
edzitron.com
Something I'm coming to realise: everybody talks about megawatts and gigawatts of data centers but there is no real consensus on what that means. Is it power? Is it compute?

Also: nobody can/will/wants to talk about what these things cost to run
justinbuist.bsky.social
Some folks were alive for the first airplane flight and the Moon landings.
raxkingisdead.bsky.social
you ever think about those real weird overlaps. like tennessee williams might have listened to the ramones
justinbuist.bsky.social
You never said it did.

I think we are very much in agreement and I apologize for any antagonistic comments on my part.
justinbuist.bsky.social
... so I cashed in on my rep and contacts. I gave it a good go anonymous;I wanted to see what it was like, to some degree, for "kids" coming into industry.

Result: You gotta know somebody more than ever since 1990.

But you can "know somebody" through the internet now so not as bad.
justinbuist.bsky.social
... it used to be if you put a job like that online you'd get 1,000 resumes total per slot. And "used to be" is 2021.

I couldn't rise to even interview level slinging out applications everywhere I was interested in. I've never seen rejection at that level anywhere in my career, even when young...
justinbuist.bsky.social
I came back into the job market after 2 years out. Good luck getting through the initial resume review these days, even using AI. I mean, I'm looking at remote data engineering work so every single applicant can use AI tools, and it's remote. Literally 1,000 resumes a day per job are flying in...
justinbuist.bsky.social
And the story is the same this go around. Inhalers are using a chemical as a propellant bad for the environment. If it gets banned for use in them for environmental reasons generics are going to go away and the exact same drugs come back with new propellants, under patent, at higher costs.
justinbuist.bsky.social
I wasn't clear. Albuterol didn't get better after the ban. The active is the same. There was no benefit to the patient other than no long exhaling an ozone depleting CFC. You couldn't measure a difference in a patient before or after the change.

Just costs 2x more when it went under patent again.
justinbuist.bsky.social
That's a good approximation.
justinbuist.bsky.social
It does when the original is illegal to produce. That's what happened with Albuterol.
justinbuist.bsky.social
I was there about 15 years ago. It sucks and I'm sorry you're feeling it.
justinbuist.bsky.social
Bikes are kind of silly. What you really want is a Vespa with a mounted 75mm rifle on it for your paratroopers.

en.m.wikipedia.org/wiki/Vespa_1...
Vespa 150 TAP - Wikipedia
en.m.wikipedia.org
justinbuist.bsky.social
It's been done before. Albuterol inhalers had to have their propellent changed for environmental reasons about maybe 20 years ago. That allowed a new patent (just expired) and so if we reformulate again, patents come back. No generics.
justinbuist.bsky.social
Azure runs more Linux than Windows, so... follow the leader?

It wouldn't surprise me if Windows becomes close to abandonware in 10 years.
justinbuist.bsky.social
Heh. I chuckled.
gracekind.net
Another victim of AI psychosis. Really sad 😔
Post from Terence Tao:

“I was able to use an extended conversation with an Al (link) to help answer a MathOverflow question (link) I had already conducted a theoretical analysis suggesting that the answer to this question was negative, but needed some numerical parameters verifying certain inequalities in order to conclusively build a counterexample.
Initially I sought to ask Al to supply Python code to search for a counterexample that I could run and adjust myself, but found that the run time was infeasible and the initial choice of parameters would have made the search doomed to failure anyway. I then switched strategies and instead engaged in a step by step conversation with the Al where it would perform heuristic calculations to locate feasible choices of parameters. Eventually, the Al was able to produce parameters which I could then verify separately (admittedly using Python code supplied by the same Al, but this was a simple 29-line program that I could visually inspect to do what was asked, and also provided numerical values in line with previous heuristic predictions).
justinbuist.bsky.social
Moderation at scale requires automation which will sometimes make mistakes. Happens on every platform.
justinbuist.bsky.social
They originally wanted to go around 1100km for the sats in polar-ish orbits. They would be covering less people when at the poles, less saturation of the bandwidth, so using fewer at a higher altitude made sense.

FCC kinda nixed it. So, ~550km it is everywhere.
justinbuist.bsky.social
At least the graveyard orbit won't decay for millions of years. I'm OK with letting the next species deal with it.
justinbuist.bsky.social
It has always been part of the plan to deorbit them on 3-5 year cycles.
justinbuist.bsky.social
In a direct reply to somebody trying to avoid "negative consequences" yes.
justinbuist.bsky.social
I really wish it would supplant the common CSV. So easy to work with, even no-code.
justinbuist.bsky.social
Fun! I was just watching something from Shane Ross on new Earth/Moon orbits akin to this.
justinbuist.bsky.social
LLM responses have a randomness to them. It's a temperature setting in proper jargon.

If you get rid of that randomness they sound like robots. And tend to get stuck in loops.