I’m guessing you weren’t around in the 90s then? Because the amount of money set on fire on stupid dotcom startups was also staggering.

The scale is very different. OpenAI needs to raise capital at a valuation far higher than any other startup in history just to keep the doors open another 18-24 months. And then continue to do so.

There’s also a very large difference between far ranging bad investments and extremely concentrated ones. The current bubble is distinctly the latter. There hasn’t really been a bubble completely dependent on massive capital investments by a handful of major players like this before.

There’s OpenAI and Anthropic (and by proxy MS/Google/Amazon). Meta is a lesser player. Musk-backed companies are pretty much teetering at the edge of also rans and there’s a huge cliff for everything after that.

It’s hard for me to imagine investors that don’t understand the technology now but getting burned by it being enthusiastic about investing in a new technology they don’t understand that promises the same things, but is totally different this time, trust me. Institutional and systemic trauma is real.

(took about 15 years because 2008 happened).

I mean, that’s kind of exactly what I’m saying? Not that it’s irrecoverable, but that losing a decade plus of progress is significant. I think the disconnect is that you don’t seem to think that’s a big deal as long as things eventually bounce back. I see that as potentially losing out on a generation worth of researchers and one of the largest opportunity costs associated with the LLM craze.

Sure, but those are largely the big tech companies you’re talking about, and research tends to come from universities and private orgs.

Well, that’s because the hyperscalers are the only ones who can afford it at this point. Altman has said ChatGPT 4 training cost in the neighborhood of $100M (largely subsidized by Microsoft). The scale of capital being set on fire in the pursuit of LLMs is just staggering. That’s why I think the failure of LLMs will have serious knock-on effects with AI research generally.

To be clear: I don’t disagree with you re: the fact that AI research will continue and will eventually recover. I just think that if the LLM bubble pops, it’s going to set things back for years because it will be much more difficult for researchers to get funded for a long time going forward. It won’t be “LLMs fail and everyone else continues on as normal,” it’s going to be “LLMs fail and have significant collateral damage on the research community.”

There is real risk that the hype cycle around LLMs will smother other research in the cradle when the bubble pops.

The hyperscalers are dumping tens of billions of dollars into infrastructure investment every single quarter right now on the promise of LLMs. If LLMs don’t turn into something with a tangible ROI, the term AI will become every bit as radioactive to investors in the future as it is lucrative right now.

Viable paths of research will become much harder to fund if investors get burned because the business model they’re funding right now doesn’t solidify beyond “trust us bro.”

Because Google holds a monopoly position and Epic doesn’t.

That said, the irony didn’t escape me either.

We also have real world examples like Alabama passing a voter ID law and then almost immediately turning around and closing DMV offices in poor, black counties, making getting an ID even more difficult for at-risk communities:

https://www.bloomberg.com/news/articles/2015-10-01/alabama-closes-dmv-offices-a-year-after-voter-id-law-kicks-in

Voter ID laws are very much about cloaking intentional disenfranchisement of legal citizens in a veil of preventing virtually non-existent voter fraud.

Clover is so beneficial that pre-WW2, grass seed mixes almost always explicitly advertised clover content. If you look up 19th or early 20th century catalogs, etc, listings for grass seed will nearly always not only mention that they contain a clover mix, but tout its benefits.

As you note, it was only post-war with the creation of modern herbicides that clover stopped being the norm. There was more or less a DeBeers-style PR campaign to convince people that clover is a “weed” since it can’t survive weed killers.

The spots that our dogs have destroyed clover, they had destroyed the grass anyway. And that’s under an old magnolia tree where everything struggles anyway. The rest of the back yard is fine.

The BSG reboot really suffered from being a product of its era.

It’s when shows were first really dipping their toes into telling an overarching narrative, but writer’s rooms were still very much geared toward producing stories of the week. The result was that a lot of shows at the time would start incredibly strongly, set up a lot of really interesting premises, and then just meander along because the writers were literally making things up along the way and because there was no coherent plan.

Know how Game Of Thrones fell apart in the last couple of seasons when they outran the preplanned narrative of the books? That’s how a lot of TV ended up in the early 2000s. BSG and Lost are probably the two most prominent examples from around that time, but it was a pretty common problem as the format of TV shows was starting to change.

dork enlightenment

I have no idea how I’ve never seen this up to now, but good god that’s the perfect encapsulation of that particular sect of morons.

The entire system is deeply corrupt beyond false positives.

We know for a fact that Russia was systematically cheating testing and the grand sum of the punishment they faced for it was having to compete as “Olympic Athletes from Russia” for two years.

It also helps that their attempts to redirect back mostly serve to highlight their weird preoccupations.

Things are happening like a former Trump speechwriter posting “Emmett Till was weird” on Twitter because they can’t comprehend just how unhinged and generally weird saying something like that is to a normal person.

Or they think they’re being clever flipping the script and ranting about “boys saying they’re girls is weird.” “Why do you spend so much time obsessing over what children have in their pants? That’s really weird.”

It all puts them in a bind. If they try to defend what they’re saying as normal, it’s very clear that it isn’t. If they try to deflect with what they think is weird, it just shows how detached they are from normal reality. It’s a surprisingly effective line of attack that largely neutralizes their normal gish galloping.

I’d be in favor of more. 26 is just because I think there’s a very easy argument to make for “every circuit gets direct representation on SCOTUS” and it’s not a huge leap to go to two per circuit from there.

Increasing throughput is definitely one of the reasons I’d support doing this as well. Thanks for highlighting that since I didn’t.

The problem is that the process for amending the Constitution is heavily, structurally biased in favor of the Republicans now. The GOP would absolutely rally around this issue because it’s one of the primary things allowing them to hang on to power right now.

I don’t believe in engaging in theatrics with a zero percent chance of success when there are real, feasible steps that could be taken to make things better.

Article III only lays out there there will be a supreme court and a Chief justice and makes Congress responsible for establishing them. It does not lay out the makeup or structure of that court. The current body of 9 justices is set by federal statute and could be changed by a simple act of Congress.

Article III also explicitly states that whatever Justices are appointed hold their office as long as they maintain good behavior (I e., as long as they haven’t been impeached) and that Congress cannot reduce their pay.

Term limits are explicitly unconstitutional.

Setting the number of judges is explicitly within Congress’ constitutional powers.

Randomized panels would probably be challenged just because it’s never been tested, but the language in the Constitution re: Congress establishing the Supreme Court is vague. That said, Congress has already established inferior Federal courts that operate in this manner, so there’s precedent.

I think you’re missing the point.

As things stand now, you get cases that are tailor made to the whims of specific people because there’s a 100% chance it ends up in front of those specific people. That’s an absolutely massive problem.

The point is that you’re less likely to have cases that are specifically aimed at stroking any given individual’s brand of crazy when there’s only a ~1 in 3 chance they’ll even hear it. A panel of 9 from a pool of 26 means that you go from a 100% chance that, say, Alito and Thomas, hear a case together to around 12%. That’s a huge gamble when it takes years and a massive amount of money to get a case in front of SCOTUS.

No, it doesn’t solve all conceivable problems with the court. But it’d help address the fact that SCOTUS justices are entirely too powerful as individuals and it can be done via simple act of Congress.

Appointees should just be subject to term limits and yearly affirmation votes by members of the BAR association to renew or revoke their qualifications

Not going to happen. SCOTUS terms are life appointments constitutionally. That means you’ve gotten into amendment territory which just plain is not realistic right now.

This matches the broad strokes of the approach I favor as well.

There are 13 Federal circuits. Expand to one justice per circuit, then double that.

But the core of the approach, regardless of the exact number, is to shift to having cases heard by randomized panels of judges. The amount of power wielded by individual justices right now is just insane. Dilute it down so that the power rests with the body rather than individuals.

Further, randomizing who hears any given case would help curtail the current environment where test cases get tailored to the idiosyncracies and pet theories of individual judges.

SCOTUS should be deciding cases based on rational reading of the law, not entertaining wing nut theories that Thomas or Alito hinted at in previous decisions. That sort of nonsense becomes a lot less feasible if there’s no guarantee a case will actually end up in front of Thomas or Alito.

As somebody that’s a paying Kagi user and generally happy with the service, it is interesting seeing exactly where the tradeoffs are.

While I’d say Kagi pretty much universally returns better results for technical information or things like recipes where it deprioritizes search spam, it’s also pretty clear that there are other areas where the absence of targeting hurts results. Any type of localized results, e.g., searching for nearby restaurants or other businesses tends to be really hit or miss and I tend to fall back to Google there.

Of course, that’s because Kagi is avoiding targeting to the point where they don’t even use your general location to prioritize results. It’s an interesting balancing act and I’m not quite sure they’ve hit the sweet spot yet, at least for me personally, but I like the overall mission and the results for most searches so I’m happy with the overall experience currently.

Searches are supposed to be fast at giving you the answer you’re looking for. But that is antithetical to advertising.

And we have evidence that this is exactly why it happened, too:

https://www.wheresyoured.at/the-men-who-killed-google/

While I’d highly recommend giving either the article a read or the companion podcast a listen because Ed Zitron did some fantastic reporting on this, the tl;dr is that a couple of years ago, there was direct conflict between the search and advertising wings of Google over search query metrics.

The advertising teams wanted the metrics to go up to help juice ad numbers. The search team rightly understood that there were plenty of ways they could do so, but that it would make for a worse user experience. The advertising team won.

The head of the advertising team during this was a man named Prabhakar Raghavan. Roughly a year later, he became the head of Google Search. And the timing of all this lines up with when people started noting Google just getting worse and worse to actually use.

Oh, and the icing on the cake? Raghavan’s previous job? Head of Yahoo Search just before that business cratered to the point that Yahoo decided to just become a bing frontend.

Zitron is fond of saying that these people have names and it’s important that we know who’s making the decisions that are actively making the world of tech worse for everyone; I tend to agree.

Already happening. Required training hours were roughly doubled a couple of weeks ago effective Jan 1:

https://www.11alive.com/article/news/investigations/georgia-mandatory-law-enforcement-training-hours-for-new-recruits/85-c835bdef-3984-452b-acf0-88b22629f414

That said, this was a state trooper. GSP have long been known for a culture of cowboy recklessness and special treatment codified into law. They report up directly to the Governor and are explicitly excluded from many of the restrictions put on local police (the moniker God’s Special People has been around for decades for a reason). They are one of the few major agencies in the state that still refuses to use body cameras, for example.

Institutionally, it’s a group set up to be and that views itself as special enforcers that are above the restraints put on others. GSP is routinely involved in high speed pursuits that end in either a fatal accident or a shooting.

More training is always a good thing, but I’ll just say I was unsurprised a trooper was involved here.

Also worth noting: 2K is incredibly toxic and regular paint filter masks are useless for preventing it from getting into your lungs. It’s supposed to be used while wearing positive pressure ventilated PPE.

Probably not the best choice for redecoration on the move.