I largely agree, but I think he's right about one thing: AI reducing the demand for junior developers is temporary, and a new crop of "AI native" juniors is going to turn that around. Software is almost certainly a Jevons good, and however much AI improves development efficiency, I think we're always going to want discerning humans managing it. Right now it's mostly seniors who have the skills to adapt and take advantage of what current AI is offering, but young people who learned the profession in the presence of AI will be well positioned to do the same.
Apparently he parted with the 15k shares at some point though, so presumably whatever price those were sold for is what all 25k would have been sold for, had he received them.
It seems to me that if OP had been granted 25k shares instead of 15k, he would have sold 25k shares instead of 15k. So even aside from the statute of limitations, the damages would be something like the value of 10k shares in 1993, perhaps plus 30 years of interest.
Though it took my dad a bit longer, he managed to quit for good when I was a kid. Still died of lung disease at 80, but I have to think it would likely have been much sooner if he'd never quit. Anyway, watching him struggle with it when I was young definitely make me repulsed by cigarettes (and most other addictive substances) such that I've certainly never been tempted to try them.
Back in 2012 I switched almost entirely from bookmarking to web clipping into Evernote, which has been great. Evernote itself has gone downhill, and at some point soon I'll probably migrate my collection to Obsidian (which has its own web clipper), but I'm still a big advocate of the general practice. It is indeed great to have a searchable record of stuff you've bookmarked vs just a tree of links that you'll soon forget exist.
That's true in a sense of the overall trade deficit, but not in terms of specific trade deficits with individual countries, which seems to be Trump's focus.
So again, I think the domestic imbalance plays a larger part than this analysis gives credit. (And per my original comment, the government portion of national saving is made up of, basically, taxes minus government spending, so - while the effect is complicated because taxation also affects private spending and investment - increased taxation could be expected to reduce trade deficits.)
Basically I don't think there's an economic justification for tariffs as a solution to trade imbalances, overall or (especially) bilateral. (Bilateral trade imbalances really shouldn't be seen as a problem at all. As the saying goes, I have a trade deficit with my grocery store, but that's not a bad thing.) There are situations where they can make sense in order to strategically protect specific industries for security or long-term growth reasons, but that's a separate issue, and the key word there is strategically, something I'm not seeing a lot of.
The current round of tariff wars are all about Donald Trump and his obsession with bilateral trade imbalances. Long-term bond rates would be better addressed through the balance of production and taxation in the domestic economy.
Edit: do those downvoting disagree that a focus on bilateral trade imbalances and tariffs is an idiosyncrasy of Trump, and that we would not be seeing similar trade wars with a different president? If so, please make your case. He has repeatedly stated that that is what he feels is important, and has shared explicit motivation for broad tariffs based on facile derivations from bilateral trade imbalances. So it's pretty difficult to see it instead as a reasoned response to long-term bond rates.
Per mile driven I expect it's true, just given their greater experience than most drivers. Not that I disagree with you that many Uber drivers are not actually great at driving, but neither is the average driver in North America. I expect we just notice rideshare drivers' mistakes more, and get exposed to them at a greater rate.
I do agree that Lambert's coverage of Tesla appears to be pretty one-sided, and that the company has provided significant benefits to the world, but I also think there's a reasonable case to be made that Musk himself is a uniquely deleterious figure, even if the harm he brings isn't with his bare hands.
I would bet that llms have talked plenty of people both into and out of suicide at this point.
That nitpick aside, I think that's an excellent list. Especially being able to articulate what it does and doesn't know, or how confident it is. That's something that naively sounds pretty simple, but clearly isn't. And it's something humans aren't great at either (see: Dunning-Kruger), but so far LLMs don't even really have the capability to attempt it.
If the culture shifted such that people were skeptical of original results but excited about replication, replication would be the big news and people would want to do it. How to get there, I don't know.
But the choices are weighted based on those probabilities. This doesn't affect the weightings, only how the final weighted pseudo-random selection is made.
It's not just swapping synonyms. The way llms work is by predicting the likelihood of the next token. It's inherently probabilistic. Choices are made based on weighted random number generation, based on those probabilities. Changing how you generate the random numbers doesn't degrade the output.
LLM output, as the author acknowledges here, is already non-deterministic. Next token probabilities are set, and tokens are chosen pseudo-randomly. As I understand it, this watermark is just going to be a matter of using a known seed and algorithm to make those pseudo-random choices, such that a signature can be detected. The important thing is, it's not replacing intentional choices with random ones, it's just generating pseudo-random results differently. Quality shouldn't be affected.
Every one of them has their own particular flavour of this aggravation too. Gemini has been my standard go-to for non-coding tasks for a while, but I started to get really annoyed with a couple aspects, especially how it would end almost every response with a barely related "would you like to do this next??" tangent, regardless of my prompt to the contrary. So I've been using Claude more for regular tasks, and am now running into its brand of infuriating idiosyncrasies. I'm also hesitant to try to code too much of this out with system prompts, for fear of degrading the outputs.
The issue is that anything you put in the prompt gets considered along with all the other stuff you put in the prompt. There's no way (currently) to give instructions for how to format the output that don't also affect all of the 'reasoning' along the way. So if you tell it to do everything normally, but to end every response with "Cheers", you won't just get that; you'll get different responses than you would have otherwise.
This is pretty unfortunate, because every LLM I've used has at least one tic that I find quite annoying. It would be great to be able to eliminate them. Sometimes I do, even knowing this drawback. But there is generally a cost. (Though I don't know if "lossy" is quite right, as that implies it's always a degradation. I think it's more likely to be harmful than helpful, given the models were tuned for their default state, but it is more of a random perturbation with a slight negative bias than a strict loss.)
IMO this kind of thing adds work and stress for the best students. If I'd been faced with something like this while I was in school I would have felt the need to basically curate a parallel construction AI discussion, knowing it would be judged. I understand the theory that, in the age of AI, the prompts are more interesting than the outputs, but I'm not convinced this is the way.
Things like the Denmark approach (and/or in-person written responses) seem like a better direction. Treat take-home writing the same as we've treated take-home math questions - as useful practice (for those who choose to actually engage with it), but not a significant input to grading.
I'm not clear what point you're making about PDUs.
Re 120V outlets, while batteries don't need to supply the full peak load, they are more effective the more of it they can supply. And a home with a heat pump or baseboard heaters, electric range, EV charger, or dryer for example is going to at times significantly exceed the capacity of the 120/15 outlet. Doesn't mean the battery would be useless, just that it wouldn't be as effective at its primary purpose of load shifting as a 240V system, and therefore less economical.
I assume plug-in batteries are considerably more workable in the UK on 240V, since they could charge and deploy at double the rate of a plug-in system here. Of course 240V plugs exist in North America, but they're generally only installed for dryers, ranges, and EV chargers, so people don't tend to have one just free.