Early pg wrote about Lisp and engineers should do their own testing and commodity FreeBSD on commodity Intel was better than Oracle on Sun for starting a company.
He wrote that makers and managers needed different schedules. He wrote that math has asymmetrical upsides. He wrote that you do things that don't scale while you're in the garage.
In his wheelhouse he was best in the genre, maybe not the Balzac he fancies himself as an essayist or the painter he fancies himself at all, but the best guy to listen to if you were doing a garage band startup that involved the Internet. He was surrounded by legitimate legends like Robert Tappan Morris and Trevor Blackwell, and he wrote about things he understood.
Late Soviet Paul Graham exists as the lobbyist for Garry Tan Y-Combinator, which isn't even really prestigious anymore. As far as the signalling value goes? I'd rather have a strategic from NVIDIA before YC. I would think about YC's money if literally no one else was interested. This is "ChatGPT Tha License Dawg", "die motherfuckers die motherfuckers die" tweets tagging elected officials Y-Combinator he's defending, and the vampire companies he cites as his clean wins are suitable filleted in the rest of the thread that mine would be redundant.
And the real mile marker of a guy whose audience has exceeded his depth is that he's lecturing a room full of people about how a single operation on the iPhone calculator app can teach you more about government and economics than is apparently understood by someone who has survived eight years in Congress designed to destroy people like her, who has an Economics degree cum laude from Boston University that she got while working as a bartender to support herself and her family after her father died, a situation with no parallel in pg's life or that of anyone adjacent to him in either it's highs or lows.
I got into this business substantially because pg's writing was so motivating to twenty year old me, and for that I'm still grateful. And just like I hope Kanye gets back on his Lexipro and starts making great music again, I hope that pg goes back to his roots and starts printing great technical and startup essays again instead of spewing solid waste.
But just like I can't follow Kanye down the "death con three to the jews in hollywood" road, I can't follow pg down the "think about the billionaires and don't listen to the honors economist multiple-term congresswoman" road.
One is dramatically more offensive in it's form, one is dramatically more toxic in it's substance, because there are people who take it seriously.
No serious person acting in good faith disputes that new methods of wealth creation have started appearing at a dramatically higher rate in the last two or three hundred years than any precedent before that. Everyone, including AOC (who I agree with in this instance but am not in general a huge fan of, just to be clear, I can respect a person's credentials without blindly endorsing them), would concede that point cleanly unless they were trolling.
The second point is so ill-formed as to verge on oxymoronic when examined against either of mechanism on the ground or Bayesian prior of history. New methods of wealth creation have triggered market failures admitting new methods of wealth centralization a number of times in recent history, the Gilded Age being perhaps the poster child for this failure mode, and the Great Depression being perhaps the poster child of the magnitude of that failure mode.
The sleight of hand here is recursive, the two points are one that is trivially true followed by the shellcode that looks like a test fixture, and the shellcode is a subtle rename, `s/time bash -c 'my-command'/sudo bash -c 'my-commmand/`. It's almost reasonable, except that it grants arbitrary privileges to something that definitely shouldn't have arbitrary privileges.
In both instances, pg is smart enough to know he's arguing in bad faith.
> The second point is so ill-formed as to verge on oxymoronic
Sounds like name calling. I don’t see a rebuttal.
> am not in general a huge fan of, just to be clear, I can respect a person's credentials
Are the credentials here a Bachelors from BU? Is this an LLM?
At "Glad you agree" we've left plausible benefit of the doubt that you're arguing in good faith and so I'll bow out with one procedural grade discharge, which is that the LLM accusation is quite trivially prohibited in the stare decis of dang's rulings over the years, but still sometimes rallies downvotes, so my credentials as a human are that no LLM I'm aware of (and I work in AI) has YouTube channels of code streams.
Chain of custody on that: https://news.ycombinator.com/item?id=48511333 -> https://www.youtube.com/@b7r6-c3t so no, I'm not an LLM. I have Opus read my comments after I post them so I don't persist in trivial errors of reasoning, but I credit LLMs for their output just as I expect mine credited.