This is a personal account that I have for myself, not for someone else. So, the views expressed herein-blah-blah-blah - the usual disclaimer applies. I may refer to something as "we had this or that," with "we" referring to my past or present employers, relatives, acquaintances, imaginary friends, or anyone else. This does not mean I speak for anyone else but myself, unless otherwise stated very explicitly.
All comment texts (save for quotes) are verbalized and authored by a human. However, a machine could have proofread them, or been used as a thesaurus or a rubber ducky while writing. All the natural stupidity is my own.
NOTE: "drdaeman" is preferably spelled all-lowercase. There's no structure or meaning to this nickname.
[ my public key: https://keybase.io/drdaeman; my proof: https://keybase.io/drdaeman/sigs/JhJF8esTqUTZYBHtpsr_KCev-o6kQmS19KzV07pGL3M ]
You jest, but I’m pretty sure it’ll be a thing somewhere, it’ll take its toll, and eventually fail from its own flaws. Then, chances are, there will be some lessons learnt - although, most likely, not on the first try. But that’s just a futuristic speculation.
My point was, however, that in modern age, where we’re literally on the verge of redefining humanity, we might be forced to redefine “cheating” as well. It’s all surely starting to slowly crack at the seams for the last half a century, and the pace is only increasing. When I was a kid, electronic calculators were banned (but not the slide rule, heh), nowadays, I’ve heard, even programmable ones are becoming accepted.
I think the idea is that the chances of an LLM causing death or irreparable harm to a consistently unsupervised kid are too low to be comparable with major sources of harm (like cars) and are rather on or below the level of random casual accidents, like choking or falling, where we normally start to get practical (rather than extra cautious) about the safeguards.
If true (quite probable; there aren't that many reported cases that I found-less than 40, all-time, worldwide), then we can't meaningfully distinguish between willful negligence on Anthropic's side and a "shit can happen for any reason" class situation. Especially considering those accidents seem to tend to involve various mental health issues, particularly including preexisting suicidal ideations as a comorbidity. Probably fewer than there are cases of other bona fide people talking kids into harm (verbal abuse, dares, etc.).
And if so, I guess ethics suggests that we shouldn't assume unproven malice in such a case unless there's proof of actual intent. A suspicion is not entirely without basis, but "hurting kids for profit" feels too provocative in its implications, to the extent it starts to feel misleading.
That's another vector, and everyone considers separately. But at least here one can - hypothetically - have a BAA and zero-retention agreement with Anthropic. Which means they have moderately strong incentives to wipe your data after a fairly short while. CLIA, if I don't misremember, mandates a few years at minimum.
> Ente Photos is extremely polished and it's comparable quality to Apple photos.
If only. It can’t even upload photos any reliably (I self-host). I had it simply fail to upload anything for days (it doesn’t provide any diagnostics, gotta figure out how to build and debug it myself), with no apparent reason. That’s despite keeping app in the foreground, on a charger, for hours, with video uploads and ML features all disabled so it was supposed to focus on just the photos. Server side is fine, web-based uploads work without any issues, app just doesn’t. I haven’t figured it out yet.
That’s incorrect. E2EE means encrypted data leaves the device, stored encrypted, and server(s) have no keys to decrypt it, only your (or other) client software does.
I have a multi-region homelab cluster and I share some photos with my friends in the US and my parents in Russia. I’m auto-uploading full library (basically replacing iCloud/Google Photos) and I can share links to selected photos or albums (a reachable node will be determined by a split-view DNS). All without risks of exposing my full photo archive in case either node gets seized or otherwise compromised.
(Now, this is what I’m trying to do. I set things up, but it’s not really functional at the moment, because Ente is buggy af, and I haven’t yet learned how to rebuild and debug their iOS app.)
This makes sense, but I'm not sure it's directed at the actual issue.
There are probably some subtle bugs I can't explain in the code I wrote all by myself. I sure had a few "what was I possibly thinking when I wrote this" moments working on some old code - and that's only the bits I know about. And I sure had countless times people pointing out "hey, you got this stinky here" in a code review (which is the whole point of it). Attention lapses and brain farts sure happen. Slop can be more frequent with LLMs but it's certainly not a LLM-specific issue. They're very productive, there's a literal outbreak, and by the sheer volume shadow any The Daily WTF stories.
However, I can agree that LLM-generated code most likely has higher probability of slop. But then, a policy "a human contributor MUST fully know and understand all the contents of the submitted work, in fine detail, all the way down to every single line of contributed code and documentation" would probably address that in a more functional manner. And then the code can be from an LLM or monkeys with typewriters author had seen in his sleep. That stops to matter because author takes ownership and responsibility: "here's a recognized rational agent who swears by their work". Makes non-self-authored code require a lot more effort (unless it's a trivial change for obvious reason), but arguably even more robust than self-authored code.
That is, unless the PR authors tend lie about their knowledge - but that'll be a whole different story, where LLMs will be just a background detail.
(I'm not saying Godot should be done something different - their project, their rules, let's use that as an opportunity to watch how it goes. Just musing on the matter in general, if there's any rationally explainable merit in such policy.)
I'm not really familiar with Nietzsche (I've tried to read, but can't really break through his language - maybe someday), but I've always felt like he was a very reactionary atheist. His amor fati sentiments seem have pretty strong Christian roots, even if he replants it into a different soil.
Any interaction, with another human, machine, or even self can trigger any emotions. I most certainly was pretty pissed at LLM behavior. So I'm not sure the "talking to ChatGPT doesn't" really holds true. I can totally agree that the risk of mismatched expectation is inseparable from life (unless omniscience is possible, which I doubt), but no interaction is mandatory.
What I'm arguing is that there's no mandate how to live one's life, where to source enjoyment, and what to value. If someone wishes and can actually afford to go full incommunicado - there is no some rule that says they mustn't or even shouldn't.
We aren't messing with non-contact tribes after all - folks wanna live their own way, we leave them be. Do you want to force them into the global world?
That’s just an echo of a religious-age idea that one’s life “must” include suffering. Which is merely an attempt to answer a “why” when one feels miserable, a consoling lie from the ancient times.
But what if one isn’t supposed to? A lot of people can slogan about price, and earning and breaking sweat, and so on - but is there really anything to it, except that this is how things currently are?
There’s no “why” in the actual reality. From where we are, it just is. No one is fundamentally supposed to do or be anything. They just do and are what the circumstances dictate. And circumstances can change someday.
Why not? One can surely use math even if they have no clue about how to prove theorems. I suck at math, but I use it every day, without knowing how to advance it.
I think it might be fair to say that a proof cannot be without value if it proves something meaningful to a human, that a human can use somehow? But such proof probably doesn’t belong in a library seemingly explicitly dedicated to human-graspable proofs. Just because it violates the intent.
> You don’t think there is such a thing as “understanding” a subject?
I probably misunderstood (lol) you, because I was thinking of the phenomenal feeling of "getting it" - which is the only thing that I was able to label as "understanding" that exists in humans, but (I think) not in machine models. But when we feel insight, that's not some "real" understanding, it's a heuristics-powered illusion. That's why I wrote that.
It's quite ironic: I genuinely thought I understand what you're referring to, but that was a misunderstanding.
> Knowing the fundamental buildings blocks of that subject? [...]
Ah, in that sense - as having a mental model of something at its structural level - yes, it exists. But then computer models can understand things too - LLMs are not just giant Markov chains with pure token statistics, they build internal representations way beyond that. Look for the Othello-GPT story, it's pretty small, but quite fascinating how a model had built a world model out of just moves.
> Would this person not need “understanding” of how pixels emerge into tasteful art?
First, let me be clear that the key point there was the judgement bit. A machine lacks personhood, thus it cannot be held responsible, thus cannot make a judgement. That classic IBM memo.
For a judgement alone, just knowledge is sufficient. Seeing an irregularity is pattern matching, "those pixels look sloppy" phase doesn't yet need a "why", just a trained eye. But - yes, understanding is necessary follow-up, how to make those pixels stop raising eyebrows. Gotta not just see that e.g. "this hand has weird fingers", but also why they're weird, and then how to correct that.
> This is quite terrifying
Yes, but - IMHO - not because of how shallow some understanding might be. Competence shifts are perfectly natural, skills that are in demand remain, skills that aren't atrophy, I don't see anything too scary about that.
What's uncanny is that a lot of people indeed pass on judgement and even agency to a tool that has none. "AI" takes jobs, "AI" destroys environment, "AI" makes people zombies, "AI" steals art - that's what's really scary, that a lot of slogans put a veil in front of the actual (very much human) agents and their actions.
Bernie and AOC (which aren't DNC mainstream, but prominent) had just pushed for a moratorium on "AI data centers" with a definition that includes "that are used for the development or operation of AI models at scale" (trivially sidesteppable by "we build this GPU farm to sell to whoever bids for compute" - which is actually true), plus a bunch of fancy extras bundled in like "The government must review and approve AI products before they are released
to ensure that AI products are safe and effective.", while lacking actual definition of "AI" (given that we had "AI" systems since '50s).
Yeah, the bill has a cause - it recognizes some pain points. But then it haphazardly tries to address symptoms instead of underlying issues (environmental regulations, utility pricing, land use, job security), while pushing vaguely defined regulations that allow arbitrary application. As if misdirected measures and poorly defined laws aren't already a giant issue.
This is a personal account that I have for myself, not for someone else. So, the views expressed herein-blah-blah-blah - the usual disclaimer applies. I may refer to something as "we had this or that," with "we" referring to my past or present employers, relatives, acquaintances, imaginary friends, or anyone else. This does not mean I speak for anyone else but myself, unless otherwise stated very explicitly.
All comment texts (save for quotes) are verbalized and authored by a human. However, a machine could have proofread them, or been used as a thesaurus or a rubber ducky while writing. All the natural stupidity is my own.
NOTE: "drdaeman" is preferably spelled all-lowercase. There's no structure or meaning to this nickname.
[ my public key: https://keybase.io/drdaeman; my proof: https://keybase.io/drdaeman/sigs/JhJF8esTqUTZYBHtpsr_KCev-o6kQmS19KzV07pGL3M ]