> Bitcoinâ(TM)s supply is finite its issuance is algorithmic and transparent and no one can âoepull numbers out of thin airâ to inflate its price or supply. ...so long as the social structure and community norms that undergird it remain in place. The ETH split over DAO is instructive here.

It's stupid; but at least more honest than trying to dress 'grok' up as the master control computer.

In the MS case; it wouldn't be too surprising if that order is also the one that urgency dictates. Neither is totally unavailable on-prem only; or entirely without more-chatty-than-one-would-like behavior; but if your concern is about your dependence and Redmond's potential direct control their groupware stuff is moving faster than their OSes(at least if you have enterprise licenses and someone to handle keeping them quiet) in the direction of pure SaaS.

You'll get some nagging about how Azure Arc is definitely the cool kid's future of glorious hybrid manageability; but your ability to run Windows as though it were 15 years ago is definitely greater than your ability to run Office that way.

I suspect that this won't be the last case we see; as MS has shown comparatively little interest in backing down on the future being azure SaaS, and there's no real equivalent to some steep but temporary discounts for dealing with people who have fundamental privacy and operational control issues; while it's not terribly challenging to find a special discount that makes sticking with the status quo look cheaper than trying to do a migration.

Before the Dot Com era, startups that succeeded transitioned from growth stocks in to blue chips. They settle down, focus on becoming more efficient at executing what is now proven business mode.

But modern tech stocks are expected to act like growth stocks *forever*. When they grow to their natural potential, they begin to turn to dubious practices to generate the next tranche of growth. They undermine their services in order to squeeze a bit more revenue out of them. Or they let their successful business stagnate while the rock star founder beguiles stockholders with visions of transforming into a block chain or AI company.

Back in the early 2000s, when Amazon first transitioned from being a book store to an everything store, and they just introduced Prime membership, you used the site and thought "this thing is great." Nobody thinks that anymore; it's slower, more opaque and less reliable, cluttered with knockoffs, sponsored results, and astroturf reviews. Fake sales events with phony markdowns? Who is surprised?

In my experience it is, how effective it is is directly proportional to preexisting project complexity when the commands are run. The bigger the project, and the more parts that are interfacing together, the worse it performs. But for small, simple projects and creating frameworks, it can be amazing.

I own, but do not operate, a few IT companies that manage corporations in the $600MM-$1B receivables range.

Based on our own help desk ticket software, our clients have opened 40% fewer tickets since ChatGPT was rolled out to every desk and phone. 40%. I expect another 40% drop (total 80%) by next year as end users just manage things themselves.

I won't downsize as the tickets aren't really generating revenue as much as headaches. One of my engineers had a broken PDF file that took her 6 hours to fix, and the end user spent 6 days trying to fix it themselves with Ai.

But -- the basic stuff? Reboot your computer stuff? Email rejected because you mistyped a domain name stuff?

You don't need a human, and we would probably have outsource that stuff to India anyway next year if not for ChatGPT etc.

I'm not sure what "Building the Metaverse" is supposed to even mean anymore. Is he still obsessed with Ready Player One fantasies?

I mean, if he's just talking about generating 3d assets and the like, then maybe? AI 3d model generation is pretty useful if you don't care about every tiny detail matching up to some specific form. For example, I used an AI tool to make an image of an ancient mug with cave-art scrawled around its edges. It got the broad shapes of the model right, but had trouble with the fine engravings, making a lot of them part of the texture rather than the shape, but overall it was good enough that I just left off the engravings, had it generate a mug without them, then re-applied them with a displacement map. It got all the cracks and weathering and such on the mug really nice, and the print came out great after post-processing (cold-cast bronze + patina & polishing).

(I ended up switching from cave art to Linear A, because I also plan to at some point make a Linear B mug so that I can randomly offer guests one of the two mugs, have them rate it, and thus conduct Linear A-B Testing)

I obviously don't expect better from these sorts of people; but I'm honestly puzzled as to why they would turn the screws so quickly and blatantly despite having gone to all the trouble of a reshuffle and a new lineup and some spiel about being likeable rather than Alexa just being something that you sort of poke at because Prime members were given a free surveillance puck with some offer one time.

Is Panay one of those abhuman lunatics who genuinely thinks that the only objection to relentless advertising is that it isn't "relevant" or "engaging" enough? Does he have a scorpion nature that leads him to knowingly doom his own product just because that's what he is? Is he just a figurehead who got to choose the case plastics colors and smile on stage; but some adtech business unit calls all the shots?

I'd fully expect this sort of thing to betray you; but only after enough of a honeymoon period for people to be pleasantly surprised by the behavior of the launch units so that there is actually enough of an install base to betray.

It's a cute "attack", but not practical in the real world. The "poison" stands out like a sore thumb on your loss graphs.

It's so easy to get tempted into feature bloat these days. You need a microcontroller for some simple set of features, like doing PWM control on a fan and handling a rotary switch, so you get something like a Seeed Studio XIAO ESP32S3 that's the size of a thumbnail and costs like $10, but then all of the sudden you have way more processing, memory capacity, pins, etc than you need, and oh hey, you now have USB, Bluetooth, and WiFi, and surely you should at least do SOMETHING with them, right? But the hey, for just a little bit of extra cost you could upgrade to a XIAO ESP32S3 Sense, and now you have a camera, microphone, and SD card, so you can do live video streaming, voice activation, gesture recognition... .... it really creeps up on you, because there's so much functionality in cheap, small packages today.

The irony though is that nobody really seems to bundle together everything one needs. Like, could we maybe have such a controller that also has builtin MOSFETs, USB + USB PD charging, BMS (1S-6S) functionality, and maybe a couple thermocouple sensors? Because most small devices need all these basic features, and it's way more cost, space, weight and effort to integrate separate components for all of them. The best I've found is a (bit overbuilt) card that has USB + USB PD (actually 2 of each, and reverse charging support), BMS support (1-5S), one thermocouple sensor, and a small charging display - but no processor or MOSFETs.

It sure is a good thing that 'AI' companies are notoriously discerning and selective about their training inputs and not doing something risky like battering on anything with an IP address and an ability to emit text in the desperate search for more; so this should be a purely theoretical concern.

Snark aside, I'd be very curious how viable this would be as an anti-scraper payload. Unlikely to be impossible to counter; but if the objective is mostly to increase their cost and risk when they trespass outside the bounds of robots.txt something that will just look a trifle nonsensical in places to a human but could cause real trouble if folded into a training set seems like it could be quite useful.

It can certainly be done otherwise; but it's not exactly unrelated when, in practice, a TPM is the industry standard mechanism for making a PC or PC-like system capable of cryptographically secure remote attestation; and when TPMs quite specifically mandate the features you need to do remote attestation rather than just the ones you would need to seal locally created secrets to a particular expected boot state. They are certainly can do that, and it's presently the most common use case; but locking down remote attestation was not some sort of accidental side effect of the design.

There's no doubt that AI is developing into a useful tool -- for people who understand its limitations and how long it is going to take to work the bugs out. But people have a long track record of getting burned by not understanding the gap between promise and delivery and, in retrospect, missing the point.

I think we should take a lesson from the history of the dot com boom and following bust. A lot of people got burned by their foolish enthusiasm, but in the end the promise was delivered, and then some. People just got the timescale for delivering profits wrong, and in any case their plans for getting there were remarkably unimaginative, e.g., take a bricks and mortar business like pet supplies and do exactly that on the Internet. They by in large completely missed all the *new* ways of making money ubiquitous global network access created.

I think in the case of AI, everybody knows a crash is coming. In fact they're planning on it. Nobody expects there to be hundreds or even dozens of major competitors in twenty years. They expect there to be one winner, an Amazon-level giant, with maybe a handful of also-rans subsisting off the big winner's scraps; tolerated because they at least in theory provide a legal shield to anti-trust actions.

And in this winner-take-all scenario, they're hoping to be Jeff Bezos -- only far, far more so. Bezos owns about 40% of online retail transactions. If AI delivers on its commercial promise, being the Jeff Bezos of *that* will be like owning 40% of the labor market. Assuming, as seems likely, that the winning enterprise is largely unencumbered by regulation and anti-trust restrictions, the person behind it will become the richest, and therefore the most powerful person in history. That's what these tech bros are playing for -- the rest of us are just along for the ride.

The place where TPMs potentially get toothy is remote attestation. As a purely local matter having your boot path determined to be what you think it is/should be is very useful; but, by design, you can also request that from a remote host. Again, super useful if you are dealing with a nasty secure orchestration problem(Google has a neat writeup of how they use it); but also the sort of thing that is potentially tempting for a relying party to use as part of authentication decisions.

We've seen hints at related issues on the Android side; where hardware attestation API or 'Play Integrity' API demands are made by some applications that block 3rd party ROMs, even if the boot sequence is entirely as expected(and even if the 3rd party ROM is almost certainly in much better shape than the first party one; eg. Graphene vs. some out-of-support entry level Samsung); which has chilled 3rd party ROMs considerably.

If relying parties who are important(ISPs, banks, etc.) do start demanding attestation the situation in practice becomes a great deal more restrictive.

I spend quite a bit on Discord server services, but I'm out.

GFY, Discord and governments, for mandating this bullshit nonsense.