I was expecting something about how to protect your consciousness from (or during) AI use, but I got a short 200 word note rehashing common sentiments about AI. I guess it’s not wrong, it’s just not very interesting.
Yeah if found it slightly ironic that an argument against using AI is made as an empty social media-style post. Ironically AI could have written a better one.
It is very interesting because it tackles things people love to forget when using AI.
A little over a decade ago it was scandal on how big tech companies are using peoples data, now people give it knowingly to them via all kinds of bullshit apps. So you have to repeat the obvious over and over, and even then it wont click for many.
I still feel "weird" trying to reason about GenAI content or looking at GenAI pictures sometimes. Some of it is so off-putting in a my-brain-struggles-to-make-sense-of-it way.
The question for me is, assuming that other companies are more humane and are avoiding laying people off before Christmas, what does January look like?
It used to be true, but companies seem deadset on demoralizing their workforce as much as possible.
They've raised prices as much as consumers will bear.
We're getting near the physical limits of how efficient things can get in many sectors.
If P/Es are to remain this high (and they have to for the rich to remain this rich), the profits must continue to grow far in excess of the total economy.
Companies, especially large ones, generally don't think like that. Q1 can be an active quarter for layoffs (Q1 2023 was probably the most active layoff quarter since COVID), but generally, they happen when they happen (Q2-2025 was more active than Q1-2025).
One very real factor that actually does act as a gravitational pull for doing layoffs in Oct/Nov Q4 is insurance elections. Waiting until January can create a lot of complicated paperwork and cost more. There is also very much a sense that doing layoffs before the holidays, while giving severance pay, is more humane than waiting, because they can "spend time with family" (real words I've heard stated).
I can see the logic. Layoffs are always terrible. But if I am getting laid off anyway, I would prefer to know about it before I spend a whole bunch of money during holidays.
Violent crime incidents in Seattle police’s M sectors
(the downtown core) declined 36% between June–August 2025
compared to the same period in 2024.
This was the lowest number of incidents for the summer
period since 2017.
To be honest, the core of Penrose’s idea is pretty stupid. That we can understand mathematics despite incompleteness theorem being a thing, therefore our brains use quantum effects allowing us to understand it. Instead of just saying, you know, we use a heuristic instead and just guess that it’s true. I’m pretty sure a classical system can do that.
I'm sure if you email him explaining how stupid he is he'll send you his Nobel prize.
Less flippantly, Penrose has always been extremely clear about which things he's sure of, such as that human intelligence involves processes that algorithms cannot emulate, and which things he puts forward as speculative ideas that might
help answer the questions he has raised. His ideas about quantum mechanical processes in the brain are very much on the speculative side, and after a career like his I think he has more than earned the right to explore those speculations.
It sounds like you probably would disagree with his assumptions about human perception of mathematical truth, and it's perfectly valid to do so. Nothing about your comment suggests you've made any attempt to understand them, though.
I want to ignore the flame fest developing here. But, in case you are interested in hearing a doubter's perspective, I'll try to express one view. I am not an expert on Penrose's ideas, but see this as a common feature in how others try to sell his work.
Starting with "things he's sure of, such as that human intelligence involves processes that algorithms cannot emulate" as a premise makes the whole thing an exercise in Begging the Question when you try to apply it to explain why an AI won't work.
"That human intelligence involves processes that algorithms cannot emulate" is the conclusion of his argument. The premise could be summed up as something like "humans have complete, correct perception of mathematical truth", although there is a lot of discussion of in what sense it is "complete" and "correct" as, of course, he isn't arguing that any mathematician is omniscient or incapable of making a mistake.
Linking those two is really the contribution of the argument. You can reject both or accept both (as I've said elsewhere I don't think it's conclusively decided, though I know which way my preferences lie), but you can't accept the premise and reject the conclusion.
Hmm, I am less than certain this isn't still begging the question, just with different phrasing. I.e. I see how they are "linked" to the point they seem almost tautologically the same rather than a deductive sequence.
And you do it again, you apologise while insulting me. When challenged you refuse to defend the points you brought up, so that you can pretend to be right rather than be proved wrong. Incompleteness theorem is where the idea came from, but you don’t want to discuss that, you just want to drop the name, condescend to people and run away.
Here are the substantive things you've said so far (i.e. the bits that aren't calling things "stupid" and taking umbridge at imagined slights):
1. You think that instead of actually perceiving mathematical truth we use heuristics and "just guess that it's true". This, as I've already said, is a valid viewpoint. You disagree with one of Penrose' assumptions. I don't think you're right but there is certainly no hard proof available that you're not. It's something that (for now, at least) it's possible to agree to disagree on, which is why, as I said, this is a philosophical debate more than a mathematical one.
2. You strongly imply that Penrose simply didn't think of this objection. This is categorically false. He discusses it at great length in both books. (I mentioned such shallow dismissals, assuming some obvious oversight on his part, in my original comment.)
3 (In your latest reply). You think that Godel's incompleteness theorem is "where the idea came from". This is obviously true. Penrose' argument is absolutely based on Godel's theorem.
4. You think that somehow I don't agree with point 3. I have no idea where you got that idea from.
That, as far as I can see, is it. There isn't any substantive point made that I haven't already responded to in my previous replies, and I think it's now rather too late to add any and expect any sort of response.
As for communication style, you seem to think that writing in a formal tone, which I find necessary when I want to convey information clearly, is condescending and insulting, whereas dismissing things you disagree with as "stupid" on the flimsiest possible basis (and inferring dishonest motives on the part of the person you're discussing all this with) is, presumably, fine. This is another point on which we will have to agree to disagree.
The only thing I can come up with is that compressing several hundred million years of natural selection of animal nervous systems into another form, but optimised by gradient descent instead, just takes a lot of time.
Not that we can’t get there by artificial means, but that correctly simulating the environment interactions, the sequence of progression, getting the all the details right, might take hundreds to thousands of years of compute, rather than on the order of a few months.
And it might be that you can get functionally close, but hit a dead end, and maybe hit several dead ends along the way, all of which are close but no cigar. Perhaps LLMs are one such dead end.
I don't disagree, but I think the evolution argument is a red herring. We didn't have to re-engineer horses from the ground up along evolutionary lines to get to much faster and more capable cars.
The evolution thing is kind of a red herring in that we probably don't have to artificially construct the process of evolution, though your reasoning isn't a good explanation for why the "evolution" reason is a red herring: Yeah, nature already established incomprehensibly complex organic systems in these life forms -- so we're benefiting from that. But the extent of our contribution is making some select animals mate with others. Hardly comparable to building our own replacement for some millennia of organic iteration/evolution. Luckily we probably don't actually need to do that to produce AGI.
Most arguments and discussions around AGI talk past each other about the definitions of what is wanted or expected, mostly because sentience, intelligence, consciousness are all unagreed upon definitions and therefore are undefined goals to build against.
Some people do expect AGI to be a faster horse; to be the next evolution of human intelligence that's similar to us in most respects but still "better" in some aspects. Others expect AGI to be the leap from horses to cars; the means to an end, a vehicle that takes us to new places faster, and in that case it doesn't need to resemble how we got to human intelligence at all.
True, but I think this reasoning is a category error: we were and are capable of rationally designing cars. We are not today doing the same thing with AI, we’re forced to optimize them instead. Yes, the structure that you optimize around is vitally important, but we’re still doing brute force rather than intelligent design at the end of the day. It’s not comparing like with like.
> correctly simulating the environment interactions, the sequence of progression, getting the all the details right, might take hundreds to thousands of years of compute
Who says we have to do that? Just because something was originally produced by natural process X, that doesn't mean that exhaustively retracing our way through process X is the only way to get there.
Who says that we don’t? The point is that the bounds on the question are completely unknown, and we operate on the assumption that the compute time is relatively short. Do we have any empirical basis for this? I think we do not.
The overwhelming majority of animal species never developed (what we would consider) language processing capabilities. So agi doesn't seem like something that evolution is particularly good at producing; more an emergent trait, eventually appearing in things designed simply to not die for long enough to reproduce...
Define "animal species", if you mean vertebrates, you might be surprised by the modern ethological literature. If you mean to exclude non-vertebrates ... you might be surprised by the ethological literature too.
If you just mean majority of spp, you'd be correct, simply because most are single celled. Though debate is possible when we talk about forms of chemical signalling.
Yeah, it's tricky to talk about in the span of a comment. I work on Things Involving Animals - animals provide an excellent counter-current to discussion around AGI, in numerous ways.
One interesting parallel was the gradual redefinition of language over the course of the 20th century to exclude animals as their capabilities became more obvious. So, when I say 'language processing capacities', I mean it roughly in the sense of Chomsky-era definitions, after the goal posts had been thoroughly moved away from much more inclusive definitions.
Likewise, we've been steadily moving the bar on what counts as 'intelligence', both for animals and machines. Over the last couple decades the study of animal intelligence has been more inclusive, IMO, and recognize intelligence as capabilities within the specific sensorium and survival context of the particular species. Our study of artificial intelligence are still very crude by comparison, and are still in the 'move the goalposts so that humans stay special' stage of development...
Since you seem to know about the subject, how are these not immediately found and shut down? It seems like the messages they send could be traced to the sims physical location, and having a massive cluster of thousands of sims just sitting in an apartment also seems like an obvious giveaway. And there’s all the traceability required to rent the locations and buy the equipment. It seems like bothering with this is just asking to get caught.
Well, they did get caught. But for that to happen immediately would require a detection method that can point out the presence of a farm with only a few samples. SIMs don't know their 'physical location' and triangulation of signals in these bands in the urban environment is non trivial.
Whoever did this likely isn't all that happy that their carefully created infra was used to harass officials, which most likely is the single reason this operation got uncovered in the first place. If it would have just been used for low level crime who knows how long they could have continued to do this.
Note that these are not unique to NYC or even to the United States, they've been found in other countries as well, the UK has now criminalized possession or operation of these (but the fines are so low that I don't think it will make much difference).
> SIMs don't know their 'physical location' and triangulation of signals in these bands in the urban environment is non trivial.
IIRC modern cell towers use cool tricks to send stuff for a particular phone to only where that phone is so they can send more total data. Can this not be turned into a precomputed map by taking a test phone everywhere and seeing what settings the tower picks to talk to it?
A portable spectrum analyzer. A high concentration of phones like this would light up the spectrum when used with a directional wand.
Portable spectrum analyzers are regularly used to identify interference in urban environments. Even a damaged cable coax line on the street can interfere with cellular signals.
With 5g and beamforming and mimo and decent bts software(Ericsson or Hua) you can pinpoint the given phone very accurately (within 20m in urban settings) - without any triangulation, as you know the cell tower sector :) Guess what: you can also measure the azimuth within 0.1 degree, so you could have SOME data at where to look.
FYI: That was available back in 2022 as standard. Now it could be even better. :P
I've already narrowed it down to four buildings for you, so we can consider that all of those methods worked. What is your next move?
I'm not saying it can't be done, clearly it can be done otherwise this article wouldn't exist. But it is not quite as easy as pointing a magic wand (aka an antenna) at a highrise and saying '14th floor, apartment on the North-West corner', though that would obviously make for good cinema.
> I've already narrowed it down to four buildings for you, so we can consider that all of those methods worked. What is your next move?
Subpoena the power, water & gas company, and look at apartments that have unusual power usage, coupled with almost zero water & gas usage. Especially look at apartments that don't have a spike in power usage in the morning & evening that corresponds to people having a regular commute.
I'm not sure how much power this equipment draws at idle - I'm assuming it's more idle at night, no need to send scammy SMS messages at 3am Eastern - but I'd wager you could track that.
Granted, it's not fast, but depending on how quickly the companies bend over backward for such a request & how good your interns are at using Excel, you might be able to get this done before sundown.
Maybe in a city like NYC with old apartments you could do that. It’s common for newer LEED buildings to use heat pumps and collective water/sewer billing. Power maybe but WFH is common these days too. And then you’d have to convince a judge that you’ve got something narrow enough.
This inspired me to find this catalog, thank you for mentioning it!
For those who have not seen it before, Waterwitch is on page 43 of the 2013 catalog here [1], and is described as "Hand held finishing tool used for geolocating targeted handsets in the field". It did seem to require, if I'm reading right, that the target be connected to a malicious GSM router called "Typhon" (page 42).
If even a fraction of those antennas are transmitting at any given time, which you can arrange simply by having the network poll them, all you need to do is wander up and down the hall with a TinySA or something similar. It will be almost ridiculously obvious where all the RF racket is coming from.
Even before doing that, a handheld Yagi in the parking lot will easily narrow it down to a couple of floors in a specific quadrant of the building.
Yeah modern cellular and WiFi modems use multiple antenna and beam forming to allow multiple same frequency connections to occur, without interference.
But when people think of beam forming as “pointing a beam at a phone” that’s kinda thinking of the problem backwards. Modems beam form by looking at the various bits of signal delay coming down multiple antenna, and computing a transform function that will effectively result in the signal it sends mimicking those delays and thus forming a beam pointing in the opposite direction of the incoming signal.
But the modem has no idea what physical direction that beam is pointing in, and doesn’t care. It just know how to analyse an incoming signal to effectively mask the inputs from different antenna in order to extract a very weak signal, by taking advantage of constructive interference between a signal received on multiple antenna, and in turn invert that function to create an equivalently strong constructive interference pattern at the source of the signal when replying.
Most important the modem has no idea what the actual signal path was, it could have bounced of several buildings, been channeled by some random bit of metal acting as a wave guide, or any other manner of funky interference that literally any physical object creates. All it knows is that is a viable signal path must exist (because it received something), and it can compute a function to send a return signal back down the same path. But it’s very hard to turn that abstract signal path function the modem understands, into an actual physical direction. Not without doing a load of extra calibration and sampling work to understand exactly how all the antenna the modem uses interact with each other, which nobody does, because that information won’t improve the cell towers performance.
Indeed. The output of the beamforming algorithm is something like four (complex) numbers that you use to tell which of your radios to shout the loudest (and with what delay), which magically makes the signal become the strongest possible at wherever the device was last heard. And at an infinite amount of other places.
If you have MIMO, i.e., multiple signal streams, it will be more like an 4x4 matrix instead (how loud should radio X shout signal Y), and you'll not only optimize for “signal 1 should be the loudest possible at receiver 1” but _also_ “signal 1 should be at the _most quiet_ possible at receiver 2”.
The fact that cheap consumer devices are able to do this fairly reliably (one could even say it's pedestrian) at near-gigabit speeds says something about how insane our level of technology is.
I think it is the same kind of magic thinking about 5G that causes people to believe that those base stations somehow mysteriously know to within a couple of feet where a handset is located. That's just not how it works, at all. At best you could say that the interference pattern caused by a particular engagement of the radios has a local peak that - hopefully - coincides with the location of a particular handset. But there are countless such interference patterns and no single one will stand out to say 'that's the one', besides the impossibility of actually calculating the patterns because of the lack of knowledge about the environment.
It's also amusing to see lots of people state with great authority how simple it is to track down a transmitter, when in fact they've probably never so much as participated in a fox hunt, which can get quite interesting at higher frequencies and when not in open field.
Thank you for this. I feel like I’ve finally gone from the 0% understanding of how beamforming works, where I’ve been for a decade, to “some basic appreciation for the concept”!
Yes, but they don't know physical location, just a complex number matrix of how each receiver perceives each transmitter, which is inverted to determine how to transmit to optimize that receiver's reception. They don't first determine location and then optimise based on location - they optimise based directly on how the radio waves propagate.
They could probably be quickly found if someone is looking for them, but carriers don't necessarily care that much about these. Add a couple layers of indirection with MVNOs and there's a lot of meh to spread around.
If the reporting around this is accurate, sounds like someone(s) was swatting through these, which brought the attention needed to find this group.
> Since you seem to know about the subject, how are these not immediately found and shut down?
Because - depending on cell tower coverage and the antennas installed on it - the degree of precision is far too low to be useful. In rural installations and the worst case, aka a tower with a dipole antenna on a mountaintop, at 900 MHz the coverage will be around 35 km. Segmented antennas just limit the section of the circle where the endpoints are. In suburban areas, coverage is usually 10-20 km, and urban areas it's 5km and less.
Now you know which cell and cell section the user is in... but to actually pinpoint the user? That takes some more work. First, you need a few more towers that the user can reach for triangulation - the more the better - but if the operator of such a setup is even remotely clever and the hardware/firmware supports it, they will have locked the devices to only connect to a single tower (you can see a map at [1] that shows the IDs). If the operator didn't do that but the site is too remote to achieve triangulation, you might need to drive around in a van and use an IMSI catcher, aka a phone tower emulator, and hope that eventually the site's devices register at it. That, however, is a lot of awful work, and is often not legal for police authorities, only for secret services.
Now you might ask yourself, what about 911, how can they locate callers precisely? The thing is... it depends. Landlines and VoIP lines are usually mapped to a specific address (which is why VoIP providers give you an explicit warning that, if you do not keep that record up to date, 911 calls will be misrouted!), so that's trivial. Mobile phone callers however, until a few years ago the degree of precision was exactly what I just described - it completely depended on celltower coverage, with the only caveat that a phone will connect to another operator if it shows a stronger signal for 911 calls. Only then, Android introduced Emergency Location Service [2] and Apple introduced Hybridized Emergency Location [3] - these work with the sensors on the phone, most notably GPS/GLONASS/Beidou, but also SSIDs of nearby WiFi APs and specific Bluetooth beacons. Downside of that is, of course, the 911 dispatch needs an integration with Apple and Google's services, users can disable it for privacy reasons, and older phones won't have anything - so in these cases, 911 dispatchers are straight out of luck and again reduced to the above range of precision.
