The available information isn't terribly specific; but on the surface it doesn't sound any deeper into application internals than it already does.

You can certainly go deeper into etw trace hell; but Task Manager knows how much CPU time you are eating, your memory consumption metrics, your network traffic, and has a (currently less granular) look at your consumption of GPU time. It also having a view of consumed NPU time seems totally analogous for a new flavor of coprocessor. Only one that might involve gory poking at the application(though I'd assume would be more logically pulled from the GPU driver) would be distinguishing "GPU Neural Engine" from GPU utilization generally.

Investors are chasing a company that achieves some kind of AI singularity. Let's set aside the fact that there's no reason to believe there is anything but diminishing marginal returns by making marginal refinements to current frontier models. Let's imagine someone hits the jackpot and gets, not even AGI, but a system that's as far ahead of today's frontier model are ahead of 2020's GPT 2.0 in performance.

Globally AI revenues are 150 billion, against a cumulative burn rate of 450 billion. A model that is a generation ahead of others would almost certainly capture the lion's share of that revenue.

If AGI magically appears as a Sam Altman has promised investors it will, a hundred million is way too low. Add, maybe, another zero to the revenues.

Conservatively, a safer assumption is that frontier models will get marginally better based on refinements in training and reinforcement and the other bits and bobs that go into these systems. The nobody is winning the lion's share of anything, at least overnight. But you have to define "safe". By "safe" I mean unlikely to lose money. But some investors are clearly defining "safe" as "having the greatest chance of owning a piece of the biggest thing ever."

I'm not following this super-closely, but if Anthropic is pursuing adding multi-step model based reasoning to their system, that could be the basis of a generational leap in capability. But if that is an approach that looks like it has a chance of working, then their competitors are no doubt pursuing the same thing. In that case you'd expect the revenue pie to grow as the scope of model utilty increases, but that growth to be split among several competitors. This could credibly result in a revenue stream for some of them that is as big as the entire industry's revenue stream today. But there's going to be hell to pay on the data center impacts end of things.

Yeah.

They completely ignore panserpmia.

They just quote Isaac Asimov's writing from almost half a century ago (Europa). Ignores Enceladus, Titan, Venus, Mars; all of which have *BIOSIGNATURES* of a probability curve not too dissimilar to Europa.

They are likely a dilettante. Those folks do not do "prior art" research because they believe in the Auteur Myth.

  "An auteur is a filmmaker whose personal style and thematic focus are so distinctive that they are considered the 'author' of their films, a concept known as the auteur myth."

Real science isn't that. Real science is a rational discipline. The null hypothesis -- know what that means. Prior art -- know what that means. Mental biases -- know those, too; you'll see what you want to see, hear what you want to hear. Science is the discipline of mind that ensures you're not just spewing random nonsense. You verify what you think; that's the discipline.

As Isaac Newton said ( https://en.wikipedia.org/wiki/... ), "nani gigantum humeris insidentes."

Ignore them.

Everything seems obvious once it is explained.

Why do 12-14 steps of algebra, when calculus works in 2?

Obvious. People should know calculus.

Yet they don't.

Define Green's Theorem: https://en.wikipedia.org/wiki/... -- is obvious.

Everything is obvious *IN HINDSIGHT*.

If you are treating Windows devices as much like chromebooks as possible (autopilot deploy into S-mode; fairly draconian setup) you can get away with not too much IT; but at that point you really have to ask yourself why you are using Windows devices.

Most of the things you actually need windows for don't play all that well with that level of lockdown; and between chromebooks being cheap and defaulting to that level of lockdown without additional effort and the low end Apple stuff being nicer than wintel hardware that costs the same it's a bit tricky to make a strong case for running Windows systems as 'edgebooks'. You can; but it's not a super compelling option.

That's exactly why it's sort of disappointing to see him promoted away from actually doing stuff; and showing off a one of the less interesting "do Doom in a technically weird way" demos around.

Plenty of c-suite are just "meh, nothing of value was lost" when the suit fully engulfs the human; either they never seem to have done anything of interest or were popped directly out of some viscous spawning pool at McKinsey or whatever; but Russinovich was really good for nasty NT crevices work, so it seems a real shame for him to be vague CTO of cloud stuff. I don't begrudge him the money the promotion presumably involved; but going from actually interesting things to sending copilot to double booked meetings is kind of grim.

It would be interesting to know what the numbers look like in terms of schools that are actually doing something at least modestly Apple-specific vs. ones basically reusing chromebook-focused plans on Apple hardware.

Anything you hand out is going to be 'a platform' and normalize it for the students(not necessarily a point in its favor; people tend to resent the deliberate limitations of hardware that IT has been told to at least try to make less useful for porn and cheating); but there is a potential difference in versatility and vendor flexibility between people actually adopting some macOS stuff and people concluding that chromebooks have been getting more expensive at an alarming clip, without feeling any less cheap, so Apple is now genuinely competitive even if you are just going to slap managed chrome on there and reuse your existing, chromebook focused, IT and curriculum stuff.

I suspect that Microsoft, not unreasonably, thinks that the value of their AI slop will be enhanced if they can use an existing property to get people to classify particularly grating AI slop so that they can tune their copilot-related stuff to produce less grating or less obvious slop in the future.

Maybe someone in the org actually cares about linkedin's reputation as a hellscape for aura-farming lunatics; but I'd be shocked if the incentive that MS is actually motivated by is anything but the chance to get a dataset of text ranked by how unpleasantly bot-like it is to try to make their bots less visible.

Looks like Microsoft is trying to convince people to classify more and less distasteful bot slop for them so they can tune copilot outputs in the future.

The winning move is not to play.

IMO, we shouldn't be banning unless there are real, verifiable security concerns.

Instead, we should be using a combination of tariffs on Chinese robotics, subsidies and R&D investment at home, partnerships with emerging robotic makers like Vietnam, or partnerships with Mexico where there's a robust and growing deployment and software development industry for robotics and automation.

We cannot be an isolated autarky, so we need to treat this as an onion: Strengthen our position at home first, collaborate with markets we are already integrated (Mexico), partner with existing markets (Germany, Japan) or emerging ones (Vietnam) where IP laws are respected, and then seek protections against competitors with unfair advantages or that violate IP laws.

And all of that while luring talent from abroad while cultivating talent at home.

A ban by itself doesn't do anything to prevent competitive stagnation.

And I hope the plaintiff wins. When I used to work in retail and fast food, I ran into similar primitive managers, for whom humilitation was, apparently, a good management strategy.

But I never saw anything like this, going into the physical.

And that's the problem with many sales positions - no shade to the industry or its workers, but we tend to see the most uneducated savages climbing to management positions. Not just bad managers, but really horrible, abusive types.

Once the supply of customers flush enough to buy under more standard arrangements is saturated you are basically left between ignoring the demand for growth forever or deciding between handling...subprime...customers yourself or keeping the arrangement at arm's length.

This 'restricted mode' that Apple is vigorously denying will be used for lease enforcement suggests that the metaphorical arm is not that long and that Apple seems to be doing more than merely being willing to sell devices to whatever resellers; but it's not a huge surprise that they'd want the risk related to the debt and any shaking down handled by someone who isn't them.

It's entirely possible that they didn't actually care that much. It seems at least as plausible that the sort of people who would risk this sort of legal exposure over a non-issue were more motivated by an excuse to throw weight around and talk like a mob boss than actual concern. Nice to be able to put a hobby project on the company card.

I find it difficult to believe that 'influencers' are a good sign at a society level; but I'm a little puzzled by the degree of hostility to what is basically a degree for self-employed PR/marketing people.

As much as I'm not happy about it, PR and marketing are absolutely entire fields; and I'd assume that people doing those things whose client is themselves would also benefit from a dash of entertainment related law; contracts and accounting related to dealing with sponsor agreements; and the same sort of low-to-intermediate business stuff that routinely trips up musicians, artists, athletes, etc. who suddenly have some worthwhile amount of money/valuable copyrights/etc. but are either wholly dependent on 'agents' of variable quality and honesty or just end up getting screwed.

With the amount of whining I heard about how this version of the Odyssey was woke race treason engineered to poison western civilization I would have expected Elon to move faster than the copy cops when it showed up on his precious everything app. Is he developing a tolerance to the stimulants? Renting out datacenter space making grok sluggish?