krellor's recent activity
-
Comment on What could be Microsoft's larger game plan or agenda with CoPilot? in ~tech
-
Comment on Housing market predictions in ~finance
krellor LinkHere's a good article on why the supply and demand signals never line up for housing: We Need to Keep Building Houses, Even if No One Wants to Buy As far as what you are approved for, most lenders...Here's a good article on why the supply and demand signals never line up for housing: We Need to Keep Building Houses, Even if No One Wants to Buy
As far as what you are approved for, most lenders look at a total debt to income ratio of no more than 35-40% of gross income. Their risk is hedged because the loan is backed with collateral, which they vet much more thoroughly at the underwriting stage, which happens after you get an offer accepted. Things like the condition of the home, how much you were approved for, etc, will go into a final decision.
But yeah, it's tough. And there are many other factors, such as developers being incentivized to build more expensive homes rather than the traditional starter homes. Counties and municipalities could regulate that better by simply requiring a certain number of units per acre in certain zones. They could also offer tax incentives to develop affordable housing, or even invest in development themselves, but that means costs up front and less revenue in the future from property taxes. And most of them kowtow to existing home owners who want to see value keep going up, which is slowed by large amounts of new affordable housing.
Bear of luck!
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentI think that's a good way to look at it. We have some observational support for things like Hawking radiation, which seems to put a lifespan on black holes, and that is in tension with any model...I think that's a good way to look at it. We have some observational support for things like Hawking radiation, which seems to put a lifespan on black holes, and that is in tension with any model choices that require an infinite blow up without some new process or model that resolves things at the end of the black holes existence.
One of the most important topics I took in college years ago was the philosophy of empiricism, which I think is an important thing to balance with theoretical models. We don't want to dismiss sound modeling, but we also don't want to be dogmatic in the face of obvious tensions with observed phenomena.
Folks that insist on infinity without at least acknowledging the unknowns and the tensions are, in my mind, still trying to grapple with the concept themselves, or probably not engaged in the observational side of things at all.
Cheers!
-
Comment on Microsoft confirms Windows 11 Recall AI hardware requirements in ~tech
krellor Link ParentI was pretty clear about the need to see what actual protections are put in place. That said, all you need to do is look at Microsofts and Google's SEC fillings to see who is more motivated to...I was pretty clear about the need to see what actual protections are put in place. That said, all you need to do is look at Microsofts and Google's SEC fillings to see who is more motivated to monetize your data as a proportion of total revenue. Microsofts search and advertising revenue is 6% of total revenue compared with alphabets 80%.
I'm not saying don't pay attention, and yes, these are large companies with many firms. But Microsoft has to maintain a reputation to keep customer data private for most of their services, unlike Google, where most of their services are ad funnels.
As someone who used to manage enterprise accounts with Microsoft, Google, AWS, etc, I hated dealing with our Microsoft reps, but I will say they never played fast and loose with the contractual commitments to keep our data private.
-
Comment on Microsoft confirms Windows 11 Recall AI hardware requirements in ~tech
krellor Link ParentAs far as the privacy goes, we need to see what the actual implementation and controls are. Microsoft makes money by selling software subscriptions, unlike Google who mostly makes money...As far as the privacy goes, we need to see what the actual implementation and controls are. Microsoft makes money by selling software subscriptions, unlike Google who mostly makes money commercializing your data. Microsoft is more incentivized to actually keep your private data private. But we will see.
As far as who would benefit? If it can automatically caption training videos made with screen capture, or use a generative model to create step by step documentation from actually doing the process (building or a wiki page on how to do the thing you recorded), etc, that could be pretty neat. How many times are people taking snips as they do a task, paste it into word with some notes, and then repeat on the next step, then clean up and publish? Like a lot. Anything that get us closer to automated process documentation, captioning, etc, are potentially huge productivity wins for some professionals.
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentYou are bringing the point right down to the razor of what is known vs what is predicted. That said, I believe you are incorrect in that the event horizon is a point of infinite dilation. The...- Exemplary
You are bringing the point right down to the razor of what is known vs what is predicted.
That said, I believe you are incorrect in that the event horizon is a point of infinite dilation. The event horizon is only a coordinate singularity. The physical singularity is the actual singularity inside the event horizon. That said, the event horizon is very, very close relatively speaking to the physical singularity. It would be like picking a point very close to an asymptote. It might be a small difference but it is the difference between arbitrarily large and truly infinite.
I can already feel someone ready to disagree with my assertion above, so let me further defend it. The usual models that show a blow up at the event horizon are modeling a simplified static black hole, not a growing one. When you model a growing one, either the object "passes" the event horizon, or the horizon "envelops" the object. What is the functional difference? Mostly thousands of hours of people arguing in physics forums.
Whether proper time really extends past the event horizon, whether event horizons ever even really form, black holes vs collapsors, etc. is more of a semantic debate than a practical one.
At the end of the day the observed phenomena we call a black hole exists, new mass is incorporated into the phenomena in proper time with the one quibble of passing the horizon versus enveloped at t=o, and the phenomena has a finite lifespan in the form of evaporation. Those core facts pretty definitively rule out "infinity" with the one exception of information reaching the distant observer in their original frame of reference.
Also, the above is the exact argument I had with the editor from nature. 🙂
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor (edited )Link ParentI think there are a few things going on. We've been a little loose with specifics, as is appropriate for a casual conversation. But I think we need to be careful with the specifics if we want to...I think there are a few things going on. We've been a little loose with specifics, as is appropriate for a casual conversation. But I think we need to be careful with the specifics if we want to clarify further.
First of all, let's disclaim a few things. First, we just don't know some things yet. We have ideas and predictions, and areas of uncertainty and speculation. There will always be some physics grad student ready to wade into a conversation and say "well actually, black holes never form, those are collapsers..." At a some point there are some things we just don't know, and some distinctions that just don't help us model or understand.
That said, let's imagine a laser approaching the event horizon of a black hole, and a distant observer, with the laser emitting photons at the observer. That distant observer is maintaining, somehow, a relatively fixed position relative to the movement of the black hole through space.
From the perspective of the laser, let's say starting at 10r, it will take less than a second to cross the event horizon at 1r (which they don't experience) and then squish into the singularity.
To the outside observer, time will slow for the laser, and it will slow down as it approaches the event horizon. The light from the laser will become redshifted as the laser approaches the event horizon. Each photon will lose energy, not losing speed but reducing the frequency until it enters the radio spectrum.
The time between each photon will increase. From the lasers perspective, it is emitting just as many photons per second as normal. But as its time slows relative to the observers time, the observer will begin experiencing an increasing delay between photons.
Eventually, when the laser crosses the event horizon, there will be one last photon emitted that escapes. After that, the photons are trapped inside. But that last one will begin its long journey to the observer.
Now, in the lasers time, once it crosses the event horizon, the observer would be able to detect the increased mass of the black hole at the speed of causality, even if it appears that the laser hasn't entered it and is still emitting photons. Those photons have been crossing vast spaces to get to the observer long after the black hole has absorbed the laser in proper time, and the increased mass begun propagating through the gravitational field at the speed of causality.
This will stay that way for a long time. The observer will have to wait a finite but unknown amount of time to receive the last photon, tired and stretched from its journey. And that might happen to coincide with the final demise of the black hole through Hawking evaporation.
But there will be a last photon. That observer just won't see it until far into the future, likely as the black hole disappears, and the observer will never be able to know when the last one was received in that frame of reference.
It's all about the frame of reference, the warping of space and time, and the consequence of general relativity which is that multiple truths can exist, but some matter more than others. The laser can be in the black hole contributing to the mass of the singularity, and to the observer outside the event horizon, still be slowly firing off photons for billions of years.
People will equivocate over things like, did the laser cross the event horizon, or did the event horizon rise up to envelope the laser, etc... And maybe some of those distinctions matter some of the time. But I think what matters most is this:
- The laser rapidly crosses the event horizon in proper time.
- in proper time, the mass of the laser contributes to the singularity, increasing the mass, and thus modifying the gravitational field propagating through space, which is really the important thing about black holes.
- any change in frame of reference approaching the event horizon will coverage to the frame where the laser crossed the event horizon and contributes to the mass of the black hole.
Everything outside of that is an artifact of warping space-time. What matters is causality. Once the laser crosses the event horizon, that's all she wrote. Any observer outside in any different frame of reference will see and experience different time dilations, etc. but none of them could travel to the laser they see and grab it before it goes in. There is only one laser, smeared across many perceptions of time, but as you change your frame of reference to approach it, it all collapses to the laser crossing the event horizon. Chasing the laser will put you right on through the event horizon, and you will experience in fast forward the lasers journey.
If nothing else, the fact that black holes have a lifespan through evaporation means that nothing can truly be held at the boundary forever, or take truly infinite time to approach, barring some fundamental change in our understanding of them.
One last caveat: at one point in my career I got into quite a heated argument with an editor from nature over sensationalist claims about black holes. So my explanation is one of multiple explanations you might hear. What I would say in defense of mine, is that it furthers a conceptual understanding at the expense of hypothetical nuances that may or may not be true, and may or may not matter. But the people pushing the truly infinite approach time argument, are, in my opinion, doing it to promote the mystic of general relativity to the general population rather than to help them develop a real qualitative understanding of what is going on.
Edit: I think I cleaned up a few inconsistencies and ambiguities, but let me know if I missed something. I'm on my phone, so it's a little hard to proof for consistency across multiple paragraphs.
-
Comment on Privacy woes and autonomy, where do I go now? in ~tech
krellor Link ParentSo basically, you run your own router inside the ISP router, this controlling your own router. 😋 Doing what you set up with a site to site VPN is of course a good way to go, but still requires...So basically, you run your own router inside the ISP router, this controlling your own router. 😋
Doing what you set up with a site to site VPN is of course a good way to go, but still requires running your own router, just not the service demarcation. However, I didn't think the OP is ready for all that goes into managing their own site to site VPN endpoint on the Internet.
But good to expound on your experiences for folks who are in a similar boat!
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentSo I think I understand a possible point of confusion. As the observer, you would see things cross the event horizon and disappear. It would be a gradual dimming, and then it would disappear....So I think I understand a possible point of confusion. As the observer, you would see things cross the event horizon and disappear. It would be a gradual dimming, and then it would disappear.
Things can cross the event horizon, just only inwards. The models that blow up at the event horizon, showing an asymptotic approach are doing that due to a limitation of the model. We can fix it, and show the path something would take inwards, while still showing that nothing from the inside will ever reach an outside observer. In one of the other comment chains in this thread, I discussed those coordinate schemes.
It's a dense read, but you can get a sense of them by skimming the Wikipedia page.
Stars can become a black by either collapsing directly into one, which takes less than a second, by collapsing into a neutron star and accreting enough matter to collapse into one, or by going Nova while leaving behind a core too dense to stabilize as a neutron star. That is the case you describe.
In that case, the black hole would form and quickly form an accretion disk from the ejected matter. The matter would be pulled in, redden, dim, and then disappear through the event horizon and contribute to the mass of the singularity. As the observer, you wouldn't see the singularity, but you would see the explosion of the Nova, the formation of the spinning accretion disk, the matter disappearing, and the lensing of light from behind the disk. Eventually, you would run out of matter to sustain the accretion disk, and all light around the black hole would dim to nothing. Then the only detectable energy would be very low levels of hawking radiation. That, coupled with the changing observation of the effects of gravity on the region coupled with any lensing of light from behind the black hole would be all you would see.
Does that help?
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor (edited )Link ParentYou are correct, I think, with maybe one clarification. Like you say, to the object falling in, there is no magic line. It goes in and squish. But any signal it emits inside the event horizon...You are correct, I think, with maybe one clarification. Like you say, to the object falling in, there is no magic line. It goes in and squish. But any signal it emits inside the event horizon never leaves.
To the outside observer, the object approaches the event horizon and then disappears. There isn't some perpetual ghost emission of the object emanating forever. That is why we can't directly observe many of the black holes we've identified. All of the matter in their accretion disk has been absorbed and no more light is being emitted.
Now, an object might appear to be really slowly disappearing into the event horizon depending on dilation effects, but that is different than the object never disappearing. You would see the object redden, dim, and eventually vanish.
Does that make sense?
Edit: @Raistlin I forgot to address your main point. The reason the black holes shrink is because they emit radiation through a quantum action. You can look into Hawking radiation for the details, but that is why they shrink. That is also part of the issue in understanding them better: we don't have a unified model for gravity and quantum effects. So light never leaves, but quantum effects create particles just outside of the event horizon that in some cases are able to leave, reducing the energy of the black hole.
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentIt would probably look pretty boring for most of the time. Once the black hole finished pulling new matter towards it, its accretion disk would shrink and disappear, which is the source of the...It would probably look pretty boring for most of the time. Once the black hole finished pulling new matter towards it, its accretion disk would shrink and disappear, which is the source of the light black holes give off.
You would be able to detect changes in the black hole by observing changes in the region of space it occupies and its shrinking sphere of gravitational influence.
We don't have good models for what happens at the very end though, when the black hole gets small. Maybe it just fizzles, or maybe something more exciting happens. We just don't understand gravity at small scales very well.
I think what might be confusing is that we talk about a particle taking forever to reach the event horizon. However, that isn't really the case, so it's not as if particles are hanging out forever emitting light and other signals.
What really happens is that light that passes the event horizon takes an infinite to reach any fixed observer outside of the event horizon. The light emitted outside the event horizon travels off into space like normal, although locally it follows the distortion lines created by the black hole.
So to an observer fixed relative to the black hole, it would be a boring show once the accretion disk was finished being absorbed.
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentI took a run at your deeper question in your other reply to me. But to answer the question on coordinate systems, I think really what you are thinking of is the metrics we use to calculate...I took a run at your deeper question in your other reply to me. But to answer the question on coordinate systems, I think really what you are thinking of is the metrics we use to calculate distances. Those different metrics reflect the spacetime we are measuring. They should always be accurate to be useful, but they are not necessarily "complete." In fact, the Eddington–Finkelstein coordinates are still not complete; we need additional terms to make them complete, but that creates a few limitations in their use. But the big takeaway is that metrics are just different ways of measuring distances in different spaces.
An old and fun read is "One Two Three... Infinity" by George Gamow, who gives a very layperson motivation for the basic metrics of spacetime without all the mathematical wrapping. General Relativity from A to B by Robert Geroch is another good book that takes a much deeper look at GR.
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentIt's been a long time since I did any deep math in physics, and that was modeling fields of moving source charges, so it's not exactly in this space. That said, I think you are maybe grappling...- Exemplary
It's been a long time since I did any deep math in physics, and that was modeling fields of moving source charges, so it's not exactly in this space. That said, I think you are maybe grappling with a conceptual issue upstream of the math of general relativity. Let's start by simplifying the math and see if that helps.
You have a good grasp, I think, on special relativity. Indeed, the time dilation factor pops out when applying the Pythagorean theorem and the classic example of bouncing light off moving mirrors. Here is one example that walks through the derivation, and there are also good videos out there on it.
Even more generally, I think we agree that nothing moves faster than causality, and it just so happens that photons move at the speed of causality. Things like time, signals, and the speed of light are emergent properties from the speed of causality. When we observe something going near the speed of causality, its coordinate time moves more slowly than ours from our frame of reference. Because the speed of causality is fixed in all frames, the only thing that is left to flex is spacetime itself.
This gives rise to our notions of metrics, which is just a way of saying how we measure between two points in different geometries. To measure the distance between two points in spacetime separated by (dt, dx, dy, dz) we would use the following metric:
ds2=−dt2+dx2+dy2+dz2
Which should look pretty familiar as it is a basic extension of our distance equation to four dimensions. The left-hand side, ds, is the line element and invariant under all reference frames. Again, since all frames of reference will calculate ds the same, what is left to flex is spacetime itself (the RHS). The sign on the time unit is what gives rise to time dilation.
Quick recap. We agree that spacetime flexes, and we agree that this is because the rate of causality is fixed in all frames of reference.
Now, with general relativity, it isn't dilation caused by different velocities but by the warping of spacetime by mass. But we have the same basic simple concepts. We will have a metric that is used to measure a line element as the distance between two points separated by (dt, dx, dy, dz). All we do is a few simple translations to account for different spacetime curvatures by switching to polar coordinates and substituting different terms.
ds2=−(1−2M/r)dt2+dr2/(1−2M/r)+r2(dθ2+sin2θ dϕ2)
If you let the mass of the object warping spacetime go to 0, you can see that these are actually identical equations.
So, the math and the concepts of the different passages of time, even infinite vs 0 time, are qualitatively the same between special and general relativity. We have to use some fancier distance calculations to measure distances in general relativity with warped spacetime, just like we would measure distance differently along the wall of a curved flower vase compared to the table's flat surface.
At this point, you might object and say, but what does it mean for light to take an infinite time to leave the event horizon to reach a distant observer? You might accept the math's qualitative behaviors and geodesics but balk at what it represents.
So, let's get back to causality. Remember, the speed of light is bounded by the speed of causality. If events in spacetime are separated such that light cannot travel between the events in the time between them, then they can not be causally linked. Once a particle crosses the event horizon, it loses its ability to have a causal link to anything outside of the black hole. That must cause the time for any signal, such as light, to take an infinite time to reach an outside observer.
That's one of many ways to think about what happens when something crosses an event horizon. There are others, and I think depending on your own perspectives and thoughts, different explanations make more or less sense. I like the causality argument because it motivates a forcing behavior for why the light has to take infinite time, i.e., the speed of light is an emergent property of causality. Any event in spacetime involving the particle inside the event horizon is forever prevented from having a causal influence on anything outside, so any signal the particle can give off must take infinite time.
All the other metrics we introduce that model the behavior differently don't change the underlying behavior of the causal barrier that is an event horizon.
Anyway, I'm pretty low on sleep now and don't know if that helped. If you narrow down where you are skeptical, I'm happy to take another run at things.
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentThat the Schwarzschild coordinates blow up at the event horizon is an artifact of the coordinate system itself. You address it using Eddington–Finkelstein coordinates, which introduces an opposite...That the Schwarzschild coordinates blow up at the event horizon is an artifact of the coordinate system itself. You address it using Eddington–Finkelstein coordinates, which introduces an opposite blow-up to negative infinity at the same rate of approach, cancelling out the behavior at the event horizon. That gives you a model to chart behavior through the event horizon while still showing that as time goes to infinity we only ever see light from outside of the event horizon.
-
Comment on First proof that "plunging regions" exist around black holes in space in ~space
krellor Link ParentHawking radiation decreases the mass slowly, and particles that go past the event horizon do actually get pulled to the singularity, increasing the mass. So hawking radiation would only make black...Hawking radiation decreases the mass slowly, and particles that go past the event horizon do actually get pulled to the singularity, increasing the mass. So hawking radiation would only make black holes smaller when there is little to no incoming particles.
When people say that it takes an infinite amount of time to reach the event horizon, they mean from the perspective of an observer some distance off. The particle itself continues to experience the normal progression of time as it passes the event horizon.
The discrepancy in observed vs experienced events is due to gravitational time dilation.
-
Comment on Meryl Streep: it’s ‘hardest thing’ for men to see themselves in female characters in ~movies
krellor LinkI think perspective taking is a skill and one that needs to be nurtured and developed like any skill. I like Daniel Kahneman's description of novels as a perspective taking technology when...I think perspective taking is a skill and one that needs to be nurtured and developed like any skill. I like Daniel Kahneman's description of novels as a perspective taking technology when referring to the impacts of the printing press.
I don't have data, and I'm leery of using intuition on things like this, but there is probably some truth in the statement if for no other reason than for a long time there were so few women leads. I don't think that women necessarily identify with every role a leading man has played. In fact, I'm sure there are many types of films and portrayals of characters that women have a hard time understanding unless they have nurtured that perspective taking skill. But there were just so many movies with male leads in all genres, that as a numbers game of course there were more opportunities for women to identify with male leads.
Like she mentioned in the article, there was a particular role that caused men to identify with her portrayal. So I think this is a sign that most people have underdeveloped perspective taking skills, and that limits the type of portrayals they can identify with. I think as we are starting to see more women produced films and women led films, we are seeing that short coming in men's, as an aggregate, perspective taking. Of course, women in Hollywood have been seeing that shortcoming all along in the form of productions not being greenlit because the male executives couldn't take those perspectives. Now it's just coming out more publicly as those movies are being made.
-
Comment on What programming/technical projects have you been working on? in ~comp
krellor Link ParentThanks for the fun reply! I've definitely hacked together some dubious shell scripts in my time. As much as I've been exasperated by finding out an organizations "database" is really just a...Thanks for the fun reply! I've definitely hacked together some dubious shell scripts in my time.
As much as I've been exasperated by finding out an organizations "database" is really just a massive, fragile Excel workbook with a hot mess of VB script under the hood, I do appreciate it for what it does well. Which is basic calculations with tabular data, and in that sense it fit the bill for my Wordle solver.
To contrast the Excel with the shell script, the "user interface" part is definitely nicer for a non-technical user in that you don't need to know regex to filter possible words. In fact, it's really not any different than the web interfaces out there. It has cells in a row for correct letter correct position, a row of cells for right letter wrong position, and a row for wrong letter. As you fill the cells in, two words lists below start filtering out, one with all possible words and one with the most likely words.
However, building the behind the scenes worksheet wasn't hard but isn't pretty. You have a matrix with about 12,000 rows, one for each five letter word. Then you have about 36 columns, one for each user input field (right letter right place first position, etc).
There are formula that create a 1 in each cell when the given word in that row matches the criteria represented by that column. You then just select the rows with the most entries, and those are your elgible words. Then I made a second set to show likely suggestions of the possible words.
I have a few more staging columns, but that is the gist of it.
I suspect the boolean calculations would throw most of today's programmers who aren't used to those sorts of tricks. But having cut my teeth on low level languages, but arithmetic is pretty natural to me.
However, I agree that for a real tool, Python is the better solution. Or c#. Or pretty much anything that isn't the shell or a giant sparse matrix. But we can all agree that the worst way would be Windows batch script. 🙂
-
Comment on Privacy woes and autonomy, where do I go now? in ~tech
krellor Link ParentI would also suggest that if going the route of flashing older model consumer routers, pick up a spare or two when you can, and flash and configure the spare before swapping it into production....I would also suggest that if going the route of flashing older model consumer routers, pick up a spare or two when you can, and flash and configure the spare before swapping it into production.
Really brings down the stress level.
-
Comment on Privacy woes and autonomy, where do I go now? in ~tech
krellor Link ParentI just wanted to add to this that wherever you land on privacy, one you've considered the tradeoffs and do what is worthwhile to do, make peace with the rest. I see a lot of people really go to...I just wanted to add to this that wherever you land on privacy, one you've considered the tradeoffs and do what is worthwhile to do, make peace with the rest. I see a lot of people really go to extreme lengths, and then still agonize and spend precious mental energy on what they can't control. I understand lobbying or supporting change, but not tying yourself into knots daily about it.
-
Comment on Privacy woes and autonomy, where do I go now? in ~tech
krellor (edited )LinkYou can't really have privacy if you don't control your router. It's just too low down in the OSI model. I run my own router and firewalls on protectli hardware, force the use of Ad guard DNS on...You can't really have privacy if you don't control your router. It's just too low down in the OSI model. I run my own router and firewalls on protectli hardware, force the use of Ad guard DNS on clients with custom filter lists, segment my networks to keep my stuff separate, etc. I also use browser plugins like ublock origin, umatrix, and privacy badger.
At a certain point you need to decide what trade offs are worth what level of privacy, and realize that spouse/kids might feel differently.
I would also counsel that you should balance your concerns with a broader perspective. Take extra steps when privacy is most important, but don't worry yourself sick about it constantly. It's unfortunately something baked into a lot of technology, so focus on what you can control, mitigate, and avoid, and make peace with what risks you do accept, for your own sanity.
There are also services that do a good job of automating your data removal from brokers. Incogni does a good job cleaning up those behind the scenes brokers, and onerep focuses more on public people search sites. Both will get you added to suppression lists.
Edit: accidently made this a reply to another comment, moved it to its own top level.
Also who your ISP is probably doesn't matter as much as you think, especially if you manage your own router, use your own DNS, and use a VPN when warranted.
I'm interested in the Copilot+ computers, since it looks like we can run our own code against the NPU.
I would want to test to see how useful it is, because the MNIST data set and number classification is a pretty trivial problem these days. But could this satisfy an intermediate need for AI compute resources for researchers who don't consistently need large GPUs, but tune a few classification training algorithms as part of a deliverable on a grant?
As far as the Copilot software companion, it's for people who need an easier interface with their computer and how to ask questions.
For the Copilot recall type features, it might be helpful as a productivity time saver if it is good enough to generate documentation, caption training, create meeting notes and summaries, etc.