Yep, all the biggest from the dotcom era were companies that provided the proverbial pickaxes and shovels, and we *mostly* see that here too (Apple largely being unrelated, Google actually a bit, but not wholly in the game of actually training models, mostly the big players are mostly providing hardware or infrastructure to all these.

The big cautionary tales that everyone remembers like pets.com and webvan had silly high valuations, but no where near the heights attained by Intel, Cisco, Oracle, and Microsoft of the time. It was just so many of those dotcoms sourcing from the vendors propping up their valuations, but the vendors remained viable businesses just as they were before.

Even OpenAI, the most hyped of the hyped purely AI plays is still relatively dwarfed by Meta. OpenAI would be an example of something more akin to the intrinsically dot-com startup. So OpenAI has probably done more for the market valuation of nVidia and Microsoft than themselves.

Of course pre-pandemic, it was around $1200/ounce, and it's been just insane since the pandemic set in until now...

True, still not at the peak, but speaking of adjusting for inflation...

Cisco from 1998 to 2001 had a crazy anomalous valuation that was the biggest of the big examples of the dot-com bubble run amok. That behemoth of a company had an inflation-adjusted market cap of about a trillion dollars. Microsoft was in same ballpark, with Oracle and Intel a bit less, but still big examples of the dotcom bubble.

This time around, Google is 3.8 trillion, Meta is $1.6 trillion, Microsoft is $3.6 trillion, Amazon is $2.5 trillion, nVidia is $4.4 trililon, Apple is $4.1 Trillion....

This bubble is just massively bigger than the dotcom bubble, with just one of the big players this time being valued even adjusting for inflation more than all the big players of the dotcom era put together, and there being a fair number more of them this time. It dwarfs the 2007 bubble in these top few players alone. When this pops, it's going to be mind numbingly severe fall..

Yeah, this is one area where LLM can certainly make one side more successful. A screw up means either the attack fails, which no worse than not trying or messing up the target system, which may not be the ideal outcome, but it's not like the attacker really cared that much about the target system...

You think the companies are deliberately keeping their models from being professional grade because of some sense of social responsibility?

That is hilarious. They are pushing as hard as they can and hyping it up even more than it is capable of performing. Any shortcomings on their part is not by lack of trying or somehow holding back.

Such a shame that CVE quality is generally crap, as it's flooded with dubious 'findings' from people trying to build a resume as a security researcher. I'm not sure why you assert this is largely still done manually, reconciling with SBOM tools in my neck of the woods is pretty much automated for detecting and flagging issues because *no one* has time to deal with the gigantic volume of CVEs. Of course another problem in those SBOM tools is they have a terrible false positive rate. Trying to follow their guidance 100% may be impossible (complete misidentification) or requires significant work (SBOM tools don't do great with 'backported' fixes, and many software components don't bother with maintaining backward compatibility, so rebasing to a new version is big).

Updating software that is vulnerable is a key component, but I wager a greater general risk is how folks configure and operate credibly secure software stacks in insecure ways.

The audience for USENET and slashdot was about 400 times smaller than the people participating in broader social media. It was much harder for a critical mass of fringe ideas/susceptible people to coalesce into isolated circles when the population was just so tiny.

Australia's youth emerge as the smartest and most together in the world....

Oh should this bubble pop, it will take out a *lot* with it.

A lot of tech companies have effectively retooled themselves so they don't know how to keep being a functional business without the AI hype spending.

The level of dedication to the LLM game dwarfs the dot-com bubble, and so too will the negative consequences...

It's the boiling frog approach to revenue. Start at an attractive rate and increase it by 'no big deal' until eventually it would be a big deal.

See also, microtransactions.

Companies have learned that customers barely pay attention to the absolute costs, and just note the incrementals they incur in the moment.

What is 56 billion dollars?

BTC market cap is $1.8 Trillion....

Might not be about the popularity, the popularity is good, it's about the affordability.

The student loans were well intentioned, but just turning the money faucet on has significantly reduced practical concerns about pricing.

There are two sorts of campuses that have been *way* nicer than almost any corporate campus I've ever seen, medical and universities. In my day it was already pretty plush, and recently toured some and it's just gotten even more crazy, super large campus in the middle of some of the most expensive real estate with just amazingly nice amenities...

These easy loans started to help tackle the problem of higher ed being a *little* expensive and unfortunately made it a *lot* expensive over time. Needed to come with some regulation on the pricing side, at *least* for public universities.

Similar story on health care, by all means help people with premiums (even better would have been public option, but putting that aside), but don't just write whatever checks the insurance companies demand, regulate the health care costs.

It's worth *something*, but the price has been outpacing inflation by a wide margin for years and years.

So we have value, but the price has been running away...

Think the point would be that even if you try to opt out of AI summaries, you end up hearing someone read a script that they used AI to generate, or read comments or emails that AI generated without your awareness. Then there's a tendency to adopt speech patterns that you see in use.

So even if you refrain, you are still inundated by the content by virtue of everyone else overusing it without specifying. Even if you have a tendency to recognize AI BS a few sentences in and go away from it, you still read probably two or three sentences and may have influenced your speech a little.

I mean, sure you can mine some crypto, but the perceived value of those is essentially nothing.

The market cap for all those cited coins together is considered about $7 billion (Monero being the vast majority of that). So mining that won't do them any good to recoup expense unless they suddenly got all the crypto-bros to abandon BTC in favor of Monero (Etherium is at $380 billion, BTC is at $1.8 Trillion).