The real lesson of all these bloody triangles is that there is a hierarchy of intellectual labor with "sustained hard critical thinking" at the bottom and "spreading clever infographics on social media" at the top. I give you.. Donovan's Hierarchy™.
I would come! And bore the barman with progressively more deranged questions betraying ignorance of quantum field theory.
Think how many browser tabs you could have open at once! I don’t trust myself.
If there are no consequences and no means of enforcement for the US, in what sense is it actually a law, as opposed to “a set of guidelines”? Genuine if depressing question.
I finally had an experience that made me understand why so many great engineers are excited about coding assistants. But after 2 or 3 days of closely supervised vibe coding I became frustrated with the quality of the code. It’s a good working sketch but it needs to be carefully rewritten in place.
Thanks! Looking forward to bingeing it once I finish Emily in Vladivostok.
Someone in DoD clearly wondered: why should only the 9/11 hijackers get to use market manipulation to profit off their crimes?
Wouldn’t this admission strengthen any legal case that Anthropic makes against the DoD?
Over-falsity of confidence and asleepyness both entered our household lexicon after this scene!
Reminds me of that old skit:
News man: Henry Kissinger, tell me, what was the high point of your career?
Kissinger: Definitely in 1973, when I won the Nobel peace price for ending the war in Vietnam.
NM: And the lowest point?
HK: Oh, I would say 1975, when the war in Vietnam ended.
1 million satellites, 1 “teragram” of Alumina: these numbers are so beyond the normal units of orbital payloads that one wonders if the plan is intentionally exaggerated so it can be scaled back by 90% to allow both sides to claim victory.
It was an attempt at humor.
From www.ways-of-seeing.com/ch3 I always recall this line about a monumental painting of a swarm of nude angels: "Men of state, of business, discussed under paintings like this. When one of them felt he had been outwitted, he looked up for consolation. What he saw reminded him that he was a man."
These days they'll define any old word just because people say it. Outrageous!
Oh, “said”, duh. One of the challenges is never knowing when to give up on a tricky word!
Also that the most commonly used word in polite conversation was “arse”.
My experience of middle English, especially read aloud, is that it is surprisingly (if ploddingly) comprehensible until I hit some word like ”I-seide” that causes the wheels to completely fall of my bus.
Great essay from Hopkins on managerialism, conflicts of interest, the AI bubble, the decline of faculty governance, and how major donations paradoxically put universities in the red.
Also, since when has the Supreme Court cared about creating a “mess” when applying a principle such as money = speech or citizens may own bazookas?
“Using go fix to modernize Go code” by Alan Donovan — https://go.dev/blog/gofix
#golang
To be fair, a good painter does the same.
Fast enough, I should say. Polling will always be slower than inotify, but git status is much faster than find(1).
All FS notification packages struggle to abstract the huge variance in the underlying platforms. I want to try to make something that is simple, portable, and fast.
Yeah, a lot of clients (e.g. eglot) don’t watch the whole tree, so every time you switch git branches the server gets confused. I’m implementing a pure polling (no fsnotify) server-side watcher in gopls, stealing ideas from git status.
“Go 1.26 is released” by Carlos Amedee, on behalf of the Go team — https://go.dev/blog/go1.26
#golang
The Kardashev stuff is sci-fi nonsense. Putting GPUs in space makes no sense economically nor physically (heat dissipation). The whole thing is an obvious play to raise money for the despicable xAI by coupling it to the respectable SpaceX.
That’s a new one. I‘ve heard of anti-SLOs to ensure resilience (e.g. lets’s periodically turn off the lockserver to see who depended on it assuming too many nines of availability). But never to play managerial chicken!
It is galling to see the trillions being spent on AI to make SW developers (and others) more productive and recall the mere millions not spent on improving old fashioned source indexing tools despite immense productivity leverage.
… is a challenge. Startups (e.g. Sourcegraph) can make a better product, but they can’t give it away for free, which is the price of their competition.
I have worked on Google CodeSearch, Kythe, and GitHub’s cross-reference index. It’s not the case that those teams don’t care. Maintaining those services was always a precarious position (“cost center”) within each org despite very positive vibes from users. Scaling (large corpora, index freshness) …
