Greg's recent activity
-
Comment on How to make the world's best black shirt in ~life.style
-
Comment on How to make the world's best black shirt in ~life.style
Greg LinkI live in plain black tshirts, and I’m a big proponent in general of things being done well even if it costs more. Just ordered, looks like it’ll ship in a month or so, will report back with my...I live in plain black tshirts, and I’m a big proponent in general of things being done well even if it costs more. Just ordered, looks like it’ll ship in a month or so, will report back with my totally unscientific and arbitrary opinions on whether these are actually worth 6x as much as the bella+canvas blanks I have to replace far more often than I’d like.
-
Comment on No, artificial intelligence is not conscious in ~tech
Greg Link ParentYeah, I was mostly responding to the direct question of “where would a hypothetically conscious Claude exist, if it were conscious?” with the answer “there would be many thousands of separately...Yeah, I was mostly responding to the direct question of “where would a hypothetically conscious Claude exist, if it were conscious?” with the answer “there would be many thousands of separately conscious instances, each localised to its own server (or single-digit cluster of servers depending on VRAM demands)”.
I went into a bit more detail on the consciousness question as a whole the other day (the 404 media article there also links to this one). The tl;dr on that is I don’t think current gen LLMs are conscious (and wouldn’t expect them to get anywhere close without more work into self-modification at a bare minimum), but I also don’t find the MS Word or AoE 2 arguments especially compelling because I don’t think they really capture complexity as a key part of the question.
-
Comment on No, artificial intelligence is not conscious in ~tech
Greg Link ParentIt’s just a call center: each instance of Claude on each server (or small 2-5 node cluster) would be independently conscious if they were conscious at all, and your query will be routed to one of...It’s just a call center: each instance of Claude on each server (or small 2-5 node cluster) would be independently conscious if they were conscious at all, and your query will be routed to one of them but not necessarily the same one. The different data centres are different offices that might contain the agent handling your query, and the server racks are rows of cubicles. Claude as a whole in this context is more like TaskRabbit or Uber than a single entity, with lots of separate Claudes all operating under one coordinating piece of chat software, if you were to fully anthropomorphise it. Honestly I could even imagine one of those “virtual PA” services in the pre-LLM days giving the company a single person’s name and having every employee use that same name for a sense of continuity.
You can get into really interesting territory around swarm intelligence and diverging clones and other philosophically weird goings on that are enabled by how we can copy, modify, and distribute LLM weights but can’t do the same with biological life - but almost none of that practically applies to current LLMs as deployed. They’re just lots of independent instances on lots of separate servers.
-
Comment on Power consumption of LLM's in ~tech
Greg LinkThe scaling from 120B to 1T+ parameters likely isn’t as bad as it might look if you just extrapolate from that post, at least based on my understanding of their methodology. They’re using the same...The scaling from 120B to 1T+ parameters likely isn’t as bad as it might look if you just extrapolate from that post, at least based on my understanding of their methodology. They’re using the same 8x A100 machine for all tests - sensibly so, to get a like for like comparison - but that’s an older machine and kinda overkill for the model sizes they’re testing, so I’d guess that it’s a good baseline for comparing the models tested but less so for getting an absolute W/token value that applies to bigger models on newer GPUs with the batching and VRAM allocations tuned perfectly.
-
Comment on Nvidia announces liquid cooling system that promises to reduce electricity consumption and cut water use by up to 100% in ~tech
Greg Link ParentCustom loops for end users are always gonna be very niche, though, and the people building them tend to do a single build or upgrade every… 3-5 years, maybe? It’s deep into amateur enthusiast...Custom loops for end users are always gonna be very niche, though, and the people building them tend to do a single build or upgrade every… 3-5 years, maybe? It’s deep into amateur enthusiast territory for sure.
This kind of system is going to be more the domain and scale of commercial building plumbers, HVAC techs, maybe people who’ve done similar installations in labs or hospitals (or people trained across from those specialties to this particular type of build), so I’d imagine it’ll be pretty substantially more reliable just from specifically trained installers that do thousands per year.
-
Comment on Does generative AI have a natural limit without a major innovation? in ~comp
Greg Link ParentSorry for the very late reply, I appreciate the discussion and I missed your response on this one! I guess what I’d say is that your line of thinking reminds me strongly of biological evolution on...Sorry for the very late reply, I appreciate the discussion and I missed your response on this one!
I guess what I’d say is that your line of thinking reminds me strongly of biological evolution on long timescales, and of human societal and technological development on shorter ones. The vast majority of random genetic mutations go nowhere, and almost none will ever be a step change for a species on their own. The vast majority of people can live long, productive, fulfilling lives without ever really hitting on an idea so new that humanity as a whole hasn’t already considered it somewhere along the way - and again, of those tiny number of truly new ideas that do happen and are valuable, the number that make a large impact on their own rather than being incremental improvements are vanishingly small.
So if that’s the case for us as people, I see the occasional model making the occasional leap beyond what was known at the time it was trained to be pretty absurdly impressive, really. I see at least conceptual scope for those models being able to stack those incremental improvements autonomously if we set them on a path to doing so, especially since they can be copied and fine tuned at a scale and speed vastly higher than biology allows.
It’s all far too abstract to say with any confidence that letting loose a few million self-modifying models interacting with each other in a “society” would or wouldn’t lead to true development of the “individuals” and the system as a whole - and like I said in another comment a week or two ago, I certainly don’t have access to the billions of dollars in hardware that it’d take to test that at meaningful scale with no real pitch beyond “it’s gonna be so interesting”. But seeing that they’re capable of creating new knowledge at all is enough to prevent me from confidently betting against those capabilities…
-
Comment on Blog post: 'AI stole my face and made me a digital flesh puppet' - should I publish my life's work when extractive AI is rampant all over the internet? in ~life
Greg Link ParentI’ve been meaning to come back and give this a real reply for days, because I think you’ve hit on by far the best take I’ve seen to date on the whole messy bundle of worries and conflicts of...I’ve been meaning to come back and give this a real reply for days, because I think you’ve hit on by far the best take I’ve seen to date on the whole messy bundle of worries and conflicts of interest and frightening changes and justified concerns and unjustified concerns and genuine exploitation.
I utterly despise the basic principle of the currently leading AI companies: that you can just extract other people's work and even identity (an artist's identity is necessarily present in their work to some degree or they are not really an artist), without acquiring permission to do so and profit from it, distort it, mangle it, tarnish it, basically do with it whatever you feel like. Their products and practices reflect this deeply problematic, colonialist attitude: violate vulnerable people without care, apologise later when the damage is already done and the profits are in your pocket. Unsurprisingly, these products have enabled like-minded individuals all over the world to assault anyone with an online presence in hopes of extracting value without having to learn how to produce valuable content themselves. People like that have always existed but as the author says, AI has made their activities frictionless.
I’ve seen so, so many people - most of them genuinely worried about either real or perceived threats to the place in the world they’ve carved out for themselves so far - advocating for regressive and destructive approaches that would have consequences far beyond what they realise because they start from an understanding that what these companies are doing is bad, and then try to fit that into the tangible structure of rules and regulations they only sort of understand already.
It seems like I far too frequently find myself trying to thread the needle between “yes, these companies are evil; no, the underlying technological concept is not evil and I actually find it fascinating and exciting, even if those companies’ implementations are tainted; no, the suggestion of modifying IP law in xyz way is not viable because it would make computers themselves illegal or require selective enforcement that we know will just favour a different group of wealthy sociopaths”.
Framing it as a question of identity, accepting that the wrongness is less tangible and more philosophical but still no less real, closes that gap perfectly.
What these companies are doing is wrong, but most of it probably isn’t illegal and arguably there might not even be a sensible way to make it illegal. I don’t necessarily share your feeling that the biggest changes are likely to fizzle out, I think we’ll see a lot individual companies culled just as we did in the dot com era, but the sweeping shift remains just as the internet itself did back then. But I do still think and hope that culture as a whole can adapt around these changes and understand a bit better why certain actions are a problem, rather than drawing battle lines that only bear a cursory relationship to the outcomes most people actually want.
-
Comment on If AI is sentient then so is ‘Age of Empires II’ in ~tech
Greg Link ParentWas that Roger Penrose, by any chance? I was fortunate enough to be at one of his talks a few years back, and he dipped into theory of consciousness among a lot of other fascinating topics. I can...Was that Roger Penrose, by any chance? I was fortunate enough to be at one of his talks a few years back, and he dipped into theory of consciousness among a lot of other fascinating topics. I can certainly believe the conceptual idea that as-yet-unexplored quantum effects, and/or higher-order interactions in the structure of the electrical fields of the neurons far beyond the signal:response “circuitry” we can currently analyse could be crucial.
We’re definitely in deep “nobody really knows yet” territory, which I always find exciting! I’ll admit I find it hard to conceptualise a “brain” that could trigger these kind of interactions we don’t even yet understand without being fairly complex - even if complex in ways we can’t yet measure - but as you say, we don’t know anything for sure right now.
-
Comment on If AI is sentient then so is ‘Age of Empires II’ in ~tech
Greg Link ParentSee this is also really interesting! It’s kinda tough properly discussing something that we don’t entirely know how to define, but I was envisaging it more as a cutoff point of some kind (a...See this is also really interesting! It’s kinda tough properly discussing something that we don’t entirely know how to define, but I was envisaging it more as a cutoff point of some kind (a necessarily fuzzy one, as all things are when they hit reality, but a cutoff nonetheless) where we err on the side of caution in the grey area but largely define above the line as above the line in a binary sense.
But then I suppose I am inherently considering a gradient on some axis too (possibly a different axis?) because I have different concerns about what I’d consider ethical treatment of creatures at different points on the list. It’s a fascinating topic, that’s for sure!
-
Comment on If AI is sentient then so is ‘Age of Empires II’ in ~tech
Greg Link ParentFair, this stuff is never straightforward! I guess that’s gonna depend on how we’re defining whatever quality we’re measuring, which I was deliberately fuzzy on because I don’t have the best...Fair, this stuff is never straightforward! I guess that’s gonna depend on how we’re defining whatever quality we’re measuring, which I was deliberately fuzzy on because I don’t have the best handle on what consciousness is (do any of us, really?), let alone sentience or sapience.
It does sound like you’re on a similar page about complexity being important, though? Even if you’re putting the boundary lower than I did - maybe from a philosophical or biological difference of opinion, maybe from a semantic one on what we’re defining - the idea that there is a gradient of some kind and a reasonable cutoff point still seems to be shared.
Honestly the only thing that’s really surprised me in this thread is from @AnEarlyMartyr - the idea of putting bacteria as a stronger maybe is a slight surprise to me, not a wild one, but doing so and confidently ruling out LLMs genuinely doesn’t fit within my worldview! For the record, I’m absolutely not of the opinion that current LLMs are conscious, although I see no reason to rule out machine consciousness as a possibility in general (even if perhaps only in the far far future). I just struggle to ascribe a “maybe there’s emergent behaviour there that we don’t understand” to a biological machine like a bacterium or a fly without being forced to ascribe the same thing to a trillion parameter ball of statistical input/output processing.
-
Comment on If AI is sentient then so is ‘Age of Empires II’ in ~tech
Greg (edited )Link ParentThe paper reads strangely to me. He painstakingly and explicitly reminds us that he’s not taking a stand on the consciousness question, but then implicitly makes it pretty clear that the AoE II...The paper reads strangely to me. He painstakingly and explicitly reminds us that he’s not taking a stand on the consciousness question, but then implicitly makes it pretty clear that the AoE II implementation is intended as a gotcha to the people anthropomorphising LLMs in even the loosest ways.
Like… isn’t it pretty standard to assume consciousness/sentience/sapience is a function of complexity? Off the top of my head, with no rigour whatsoever, without even defining specifically what quality of mind I’m measuring (because even I don’t know), I’d still expect something like this to be a pretty uncontroversial take:
- Bacteria: very unlikely
- Flies: probably not
- Dogs: hmmm… maybe a bit
- Dolphins: yeah, getting there
- Chimps: quite possibly
- Humans: yes
If you managed to get >50B parameters into the AoE II implementation, I imagine the goats would start showing patterns that hit the dolphin-or-greater level on that chart and creating outputs that would get people debating the same way they do about LLMs now.
I think I get his very broad point: that the use of language pushes people to anthropomorphise more and sooner than they otherwise would. I actually agree with that point. But then he seems to see the AoE II example as a valid reductio ad absurdum when it’s actually a totally different thing (because complexity is an inherent and important part of the question), and when it probably wouldn’t be perceived as absurd at all if it were made sufficiently complex.
[Edit] Side note: not only are the replies below fascinating, they’re doing a wonderful job of underlining how happy we are to ascribe the qualities the paper talks about to biological beings without language! I think the author really did over-index on how much language is a necessity for humans to start anthropomorphising things.
-
Comment on Keir Starmer announces resignation as leader of Labour Party in the UK in ~society
Greg Link ParentI think it’s unlikely he will avoid it, to be honest. Utility nationalisation will play well with people if he manages it, especially water companies, which might be a concrete boost. Voting...I think it’s unlikely he will avoid it, to be honest. Utility nationalisation will play well with people if he manages it, especially water companies, which might be a concrete boost. Voting reform will be transformative if he’s got both the will and ability to push it through, regardless of his personal popularity, and might be one of the few ways he can really change the path the country is on.
But ultimately what I’m seeing here is maybe a slightly better state of the country going into the next election than we otherwise would’ve had, and maybe a somewhat clean slate for Labour to manage the media better even if they shouldn’t have needed it. It swings my internal barometer a few percent away from the “Reform gets in and I take that as my cue to leave” outcome, but I still see that as depressingly likely.
-
Comment on Keir Starmer announces resignation as leader of Labour Party in the UK in ~society
Greg Link ParentHard to say. Burnham is a good candidate - better than Starmer IMO, more willing to enact policies that will make an impact people can see and feel in line with Labour’s left wing union-linked...Hard to say. Burnham is a good candidate - better than Starmer IMO, more willing to enact policies that will make an impact people can see and feel in line with Labour’s left wing union-linked history, but less divisive than Corbyn (who I liked a lot, and who was treated hilariously unfairly by the press, but who did have legitimate concerns around him too). Streeting is worse than Starmer, and I sincerely hope he doesn’t mess things up by running as he indicated he might; it does at least look like we’re safe from that for now.
Burnham has been ineligible to run (he was mayor of Greater Manchester rather than a member of Parliament), and was blocked from running in the last by-election where he could have feasibly become an MP and challenged Starmer. Which led to that previously safe Labour seat switching to the Greens instead - an absolute win in my book (and if we’re honest I imagine Burnham himself wasn’t too disappointed in that one), and a fairly good microcosm of what a lot of people on the left see as the issue with Starmer: he seems to be a genuinely principled centrist, as much as that’s a thing that can make sense, but for whatever reason appears more willing to fight the moderate left within and outside his own party than to fight the far right.
All that said, Starmer was a reasonable PM. Not great, not outstanding, but competent and sensible in absurdly chaotic and difficult times. He made decisions that I disagreed with but acknowledged as principled disagreements. He made a couple that I disagreed with and thought were objectively goddamn stupid. But overall he’s done a decent job, and the fact that he’s been pressured into stepping down for no real good reason concerns me. It’s a reminder of the power of narrative control, and of which direction that is laser-focused. The broader conversation has already flipped from “Starmer is bad for the country and has to go” to “Labour are weak and fragmented because Starmer is going” as we speak.
I think and hope this will work out well, it’s a genuine chance to replace a decent PM with someone who might even be a good PM. But it’s also a perfect example of how the people who talked the world into Brexit and MAGA are still kinda running the show, even if this is a much softer example that ultimately might not go their way.
-
Comment on Watch baseball games in realtime in 8-bit view in ~sports.baseball
Greg Link Parent"The Primer was a magical thing, but it was just a mirror. It couldn't make Nell creative; it could only show her the pieces of the world until she figured out how to put them together in a way no..."The Primer was a magical thing, but it was just a mirror. It couldn't make Nell creative; it could only show her the pieces of the world until she figured out how to put them together in a way no one else had."
-
Comment on Watch baseball games in realtime in 8-bit view in ~sports.baseball
Greg Link ParentI don’t think it was necessarily your intent to come off this way, but it seems kinda patronising to say that people just aren’t aware and of course this post wouldn’t be here if they were. It...I don’t think it was necessarily your intent to come off this way, but it seems kinda patronising to say that people just aren’t aware and of course this post wouldn’t be here if they were.
It seems like a kinda fun, kinda cute project, no? A little thing that’ll give people a smile, but that probably wouldn’t have existed at all if it had to be some huge labor of love to create?
Idk, I just wouldn’t have taken a whimsical little fan project like this as an example of why AI is a problem. If anything, I see it as the opposite: “see, the tech doesn’t just make soulless slop for corporate gain, it can be a conduit for human ideas that wouldn’t otherwise be made tangible”. Whether it’s ethical to engage with tools made by huge shitty tech companies (and whether that calculation changes if you’re using an OSS model that they’re not profiting from) is a separate question - and one we probably will ultimately just disagree on - but I think there’s a lot more to it than just assuming people will automatically hate everything LLM-generated if only they know it to be the case.
-
Comment on The room the economy can’t see in ~society
Greg (edited )Link ParentSomething that really struck me years ago when I moved to a better flat in a nicer part of London was that my council tax* went down by a quite meaningful amount, even though in theory it’s tied...Something that really struck me years ago when I moved to a better flat in a nicer part of London was that my council tax* went down by a quite meaningful amount, even though in theory it’s tied to property value. Turns out the new area has more people actually paying (fewer low income exemptions, etc) and far fewer using certain services (everyone needs waste collection and street maintenance, sure, but there’s a lot less need to spend on social care, which ends up being the bulk of the budget), so the payment levels just don’t need to be set as high to balance the costs even for more expensive properties.
I guess what I’m saying is I love the kind of space you describe, and I strongly agree on the three requirements you mentioned - the tricky part with physical spaces is often balancing the high disposable income with the local accessibility to others who don’t have that income. It does happen and does work, I’ve seen it here in London, in San Francisco as you mention, and in New York too; I have no doubt people are managing it in a whole lot of cities where there’s money to do it and a reasonable overlap of people without money living side by side [edit: perhaps a function of an area having at least a subsistence-level safety net that allows them to? With that then setting the stage for volunteers like yourselves to build the next layers of Maslow’s hierarchy on top?]. But yeah, I’ve also seen a lot of Vimes Boots Theory writ large where the areas that need these services most can’t afford to run them, and the areas that could afford to don’t feel the need because they already have a private membership to a third space that almost feels like a parody of the name.
*Property tax that pays for local services. Kinda. But also not. It’s a whole can of worms in itself, but broadly speaking “local government tax covering the important physical and social infrastructure that keeps society ticking along on a mundane scale” is enough to get the point!
-
Comment on Epic Games announces Lore open-source version control system in ~tech
Greg (edited )LinkThis could actually be really useful to me… git lfs definitely is and definitely feels like a bolt on that doesn’t really solve the large files (and particularly large binary files) problem from...This could actually be really useful to me…
git lfsdefinitely is and definitely feels like a bolt on that doesn’t really solve the large files (and particularly large binary files) problem from the ground up. For a while (maybe still?) it even needed 2x the total stored data size free during a clone operation, which was particularly annoying for repos in the “small enough to fit on my laptop rather than the NAS, but big enough to stretch the available SSD space” kind of size range.dvcis great if it fits your use case, but it has some “nuke the canonical copy of remote data with seemingly innocuous local-sounding commands” bear traps hidden in there that I don’t love, and even if you protect against those it utterly breaks down in a mixed use case where you might want to individually track hundreds of thousands of small but not tiny objects (e.g. images, hundreds of kB to few MB kind of size) that are too big and too binary individually for a non-lfs git repo, and too big in aggregate to want to use lfs either. Performance tanks completely, and the solution is to use directory-level tracking which then loses file-level granularity if you need to diff a list of what’s changed, so that’s a deal breaker in a lot of situations. I did talk to the lead dev about that a year or so ago and there was some appetite for working on it, but I don’t think it’s ever hit the top of the to do list - which is fair, it’s an open source project with limited resources and their own priorities and all - but was a blocker for me.Perforce is proprietary and enterprisey as all hell, which isn’t something I want baked into the core of my projects, particularly ones I’m considering open sourcing.
Xet also caught my eye, but it does look like a kind of turbo-lfs tuned to extremely large files more than a general purpose solution. It solves some of my issues with lfs, but probably not in a significantly better way than dvc did.
So yeah, I’ll be interested to do some more reading and some playing around with this one!
[Edit] Typos
-
Comment on How funerals keep Africa poor in ~life
Greg Link ParentI think capitalism is a problem here (and in general), but so is putting ten person-years worth of resources into a funeral - whether those resources take the form of cash spending or tangible...I think capitalism is a problem here (and in general), but so is putting ten person-years worth of resources into a funeral - whether those resources take the form of cash spending or tangible shared goods in a non-capitalist society.
-
Comment on How funerals keep Africa poor in ~life
Greg Link ParentI think the context about funerals specifically and the idea of enforced destruction of wealth adds interesting texture to this though, because I can see most of the points more or less equally...I think the context about funerals specifically and the idea of enforced destruction of wealth adds interesting texture to this though, because I can see most of the points more or less equally from either a “kinship first, then strengthening economics, then loosening the more damaging expectations” or a “loosening expectations first, allowing economics to be strengthened” point of view, but I can’t square the former with the idea that funerals are sending a decade of family income up in smoke. I’d come across a few articles about the elaborate funeral culture in the past, looking at it as a social, cultural, and artistic phenomenon - and mentioning the costs as a burden - but the idea of them evolving as a deliberate destruction of wealth to maintain existing social structures is an interesting one I hadn’t heard before, and it does make a fair amount of sense to me, even though I doubt the people participating consciously see it that way.
Yeah I’d echo that, “alright” is the word. Perfectly adequate, sometimes less consistent than I’d hope, could use a bit more attention to detail, but serving the functional purpose on an item I haven’t had time to do an obsessive deep dive on - that’s why I was quite excited to see this, actually, because it seems to be the end result from someone who already has done that obsessive deep dive!
[Edit] I used to like American Apparel a lot too back in the dim and distant past, then they were hard to find (particularly outside the US) when whatever blow up caused the switch to Los Angeles Apparel, then I heard a lot of varying and sometimes conflicting stories about the CEO being creepy and weird to a degree it meaningfully affected the employees, and by that point I’d needed to find a different brand for basics anyway because of availability and I never got around to looking for them again and figuring out if they were actually a company I wanted to support or not and if so whether they were easy to find in EU/UK. As you can probably tell, clothing is one of those things that I care about to a degree, but not so much it often manages to hit the top of my overly-full to do list if I’ve found something that kinda works…