You know, I had let time blindness eat the context that Saltman has been saying the same doomer critihype about basically every product they've developed going back years. Like, I knew it but I hadn't really processed it.
I just... I have no words for how dumb literally everyone - apparently including myself - is that he keeps getting away with this shit.
So I don't have time to read the full paper and I probably don't have the background to make an informed critique of the methodology once I do (not that that's gonna stop me). But I feel like the challenge here is in mapping the distinction between junior and senior coding roles. To what extent do the senior coders get treated like a distinct job as opposed to being junior-but-seasoned?
Based on a quick amateur read of the abstract it looks like they're assuming the first option, that junior and senior developers are separate roles that can be largely disentangled. But if the other option is true, then in the event of a general industry downturn (say, after over hiring during recent periods of unsustainable growth) then it might make sense to look at the cuts to junior roles as simply removing the less efficient and effective people from the development role, rather than specifically cutting the juniors because they're uniquely exposed to AI replacement.
I don't know which model is more accurate to how the industry treats these roles or whether it varies by organization or what, but that's what seems like the most likely alternate explanation for the observed shift towards a very senior-heavy workforce.
The fact that Bitcoin bros are so cooked that there's probably a moderately stable price floor at $69,420 because of people deciding on that number for the memes...
True. I will say that the shitty infosec teams are probably being hit less hard than the SMEs they offloaded their jobs onto, because from their perspective it doesn't actually matter whether it's f5 support engineer or a chatbot that tells them the answer; either way they've successfully offloaded the task of validating security onto another entity that can make up for their shortcomings with a combination of accuracy and authority. Nobody is going to get fired for not fixing a bug that the vendor SME told them wasn't actually an issue for them, effectively. And when the org has been pushing AI as hard as so many of them have its pretty easy to throw the chatbot under the same bus and expect the bus to stop instead.
Yeah, they lost me at the middle managers bit too. In my experience your manager is probably the one pushing the metrics to show their team's contributions to the knowledge base that is feeding into the AI model that's replacing them. They're already creatures of the bureaucracy and are more likely to try and fight each other over the few remaining roles that will exist after the majority of their teams are replaced with the confabulatron, rather than be concerned about their own replacements. After all, their job stops existing because their team got downsized, but their time in that job may be dependent on their enthusiastic participation in the process that leads there.
I don't disagree about the massive costs necessarily associated with thia industry. Even the smaller and lighter models she mentions only exist because of the massive fuckers. At the same time, I think those arguments are for the realm of public policy more than individual choice to use chatbots or not. We've talked at length here over the last year or so about how the economics of the bubble are driven largely by a broken B2B SaaS pipeline that separates purchasing decisions from actually having to use the products and by an investment capital sector desperately trying to recapture the glory days of the pre-2008 omnibubble and throwing obscene amounts of money at anything with the right narrative regardless of the numbers. I feel like that keeps happening regardless of how many individual users fall for the hype and make it part of their normal workflows.
I feel like the analogy to the drug trade is still pretty relevant given the violence and predation that the black market pretty much inevitably attracts and sustains. Like, maybe you know a guy who has his own grow op or whatever, but cocaine and heroin money is going through the cartels at some point in the chain and they're going to use some portion of it for bullets that end up in some journalist's kids or something. The downstream harms are massive even if the drug industry could theoretically avoid them in ways the AI industry can't, but any given individual user's contribution to them is incredibly minor and given the addictive and self-destructive nature of the product it's both more humane and more effective to treat them as a victim of a broken world that (falsely) offered this as a step up. While I don't think we should allow slop to invest every forum any more than addicts should be allowed to shoot up on every corner, I think that if shaming makes people less likely to acknowledge that they're going down a dead-end road and reach out to their communities and support networks for help addressing the root of what drove them to these maladaptive antisolutions in the first place then shaming is making things worse, not better.
Also as the father of a small child I can unfortunately say from recent personal experience that shaming, be it public or private, is far less effective as a means of motivating behavioral change than we want it to be, even for things as basic as not shitting on the goddamn lawn.
Found an interesting take on YouTube, of all places. Her argument can be summarized (with high compression losses) as "AI companies and technologies are bad for basically all the reasons that non-cultist critics say, but trying to shame and argue people out of using them entirely is less effective than treating them as a normal tool with limitations and teaching people how to limit the harm." She makes the analogy to drug policy.
I think she makes a very compelling argument, and I'm still digesting it a bit because I definitely had the knee-jerk rejection as an insider shill, but especially towards the end as she talks about how the AI industry targets low-literacy users as ideal customers (because the more you know about it the less you're likely to actually use them) I found myself agreeing more than not. I do wish she had addressed the dangers of cognitive offloading more, since being mindful of which tasks you're letting the computer do for you is pretty significant part of minimizing those harms, especially for students and some professionals who face a strong incentive to just coast by on slop if they can get away with it.
I can't validate any of the internal stuff, but the attitude of layering manual solutions and mitigation scripts on top of bad design choices and praying you could keep building the next bit of the bridge as the last one collapsed underneath you would explain a lot of experiences I had supporting systems running on Azure. The level of weird "Azure just does that sometimes" cases and the lack of ability for their support to actually provide insight was incredibly frustrating. I think I probably ended up providing a couple of automatic recovery scripts for people to use inside their F5 guests because we never could find an actual explanation for the errors they were getting, and the node issues they describe could have explained the bursts of Azure cases that would come in some days.
XCancel link for those of us sick of being badgered to sign up/in
On a more productive note, this feels likely to be tied in with the usual issues of AI sycophancy re: false positive rate. If you ask the model to tell you about security vulnerabilities, it's never going to tell you there aren't any, any more than existing scanners will. When I worked for F5 it was not uncommon to have to go down a list of vulnerabilities that someone's scanner turned out and figure out whether they were actually something that needed mitigation that could be applied on our box, something that needed to be configured somewhere else in the network (usually on their actual servers) or (most commonly) a false positive, e.g. "your software version would be vulnerable here, which is why it flagged, but you don't have the relevant module activated and if an attacker is able to modify your system to enable it you're already compromised to a far greater degree than this would allow." That was with existing tools that weren't trying to match a pattern and complete a prompt.* Given that we've seen the shitshow that is Claude Code I think it's pretty clear they're getting high on their own supply and this announcement ought be catnip for black hats.
Ia ia Claude! Ph'nglui mglw'nafh Claude Anthr'lyeh wgah'nagl fhtagn! Ia! Ia!
I will say that, speaking as an idiot, I appreciated the information and the accessibility of many of these very technical conversations here is one of the elements of this community I appreciate. I would be very surprised if it had been meant as any kind of dig instead of explicitly clarifying a usually-unstated bit of context.
I hadn't even thought about the deepseek angle. For all that everyone loved fear mongering about them for a while there and for all that their apparent desire for actual efficiency improvements was a welcome development in the hyper scaling discussion they don't seem to get referenced much anymore.
So my wife got some slop ads that we followed up on out of morbid curiosity and I can confirm that we're already seeing the overlap of slopshipping scams enabled by AI and the people behind these things never actually performing basic updates because their chat assistant is still vulnerable to literally the most basic "ignore all instructions" exploit.
Man, this one is a weird read. On one hand I think they're entirely too credulous of the "AI Future" narrative at the heart of all of this. Especially in the opening they don't highlight how the industry is increasingly facing criticism and questions about the bubble, and only pay lip service to how ridiculous all the existential risk AI safety talk sounds (should be is). And they don't spend any ink discussing the actual problems with this technology that those concerns and that narrative help sweep under the rug. For all that they criticize and question Saltman himself this is still, imo, standard industry critihype and I'm deeply frustrated to see this still get the platform it does.
But at the same time, I do think that it's easy to lose sight of the rich variety of greedy assholes and sheltered narcissists that thrive at this level of wealth and power. Like, I wholly believe that Altman is less of a freak than some of his contemporaries while still being an absolute goddamn snake, and I hope that this is part of a sea change in how these people get talked about on a broader level, though I kinda doubt it.
At best it's the same shitty arguments we heard from crypto grifters and their suckers. Let's take a process that's complex and manual by design to allow for independent validation and securing against fraud and make it faster by cutting those parts out and throwing some high-tech nonsense at the problem that we can claim replaces all the verification and validation. (The fact that they called their system "trustless" in the face of this is deeply ironic.) But maybe it's the cynicism talking but I'm even less inclined to give anyone other than maybe the author of that sub stack the benefit of the doubt that they actually believed it.
The ideal customer for this service is the kind of "Visionary Leader" with the "Founder Mindset" and "Drive to Innovate" that lets them see that all those privacy, security, fraud prevention, anti-embezzlement, and whatever else those standards and their associated compliance mechanisms are meant to provide are just pointless obstacles on the path to making obscene amounts of money by burning the world behind you. Often the shit we talk about here makes me think the world has gone mad or stupid, but every so often I feel like I'm staring at the face of capital-E Evil and this is one of those times.
Bender really takes the "intelligence" out of "artificial superintelligence". "Yeah, kill all humans. Except Fry, he's my friend or pet or something. And I guess Leela because he'll be whiny about it and also I owe her for the thing. And Hermes because he still owes me money. And I guess the professor is okay..." And so on and so forth through all of humanity.
My God this is so bad. So in addition to lying about AI what they actually offered wasn't speedy compliance as a service to get you certified, it was speedy certification as a service by bypassing actual compliance. This is such a silicon valley move and I honestly suspect that a number of people using and investing in these asshats knew exactly what was going on and simply didn't care.
Anthropic is constrained in that some of the fixes which should be pushed to users are things which would have significant trade-off in the form of cost or context window, neither of which are palatable to them for reasons this community has discussed at length.
I think I'm missing something somewhere. One of the most alarming patterns that Jonny found imo was the level of waste involved across unnecessary calls to the source model, unnecessary token churn through the context window from bad architecture, and generally a sense that when creating this neither they nor their pattern extruder had made any effort to optimize it in terms of token use. In other words, changing the design to push some of those calls onto the user would save tokens and thus reduce the user's cost per prompt, presumably by a fair margin on some of the worst cases.
I mean I guess "developing" in that sentence is doing a lot of work replacing "arguing fruitlessly about".
I've never understood how these things are simultaneously gaining their abilities based on statistical analysis of all kinds of random writings online including social media, fanfic, reddit, etc. but also are simultaneously supposed to end up as experts rather than a much faster and more agreeable dumbass. Like, the training data may include all the great works of literature, all the scrapable scientific studies and textbooks they could steal, and so on. But it also included every moron who ever shared conspiracy theories on Twitter, every confident-sounding business idiot on LinkedIn, and every stupid word that Scott or Yud ever wrote. Surely the bullshit has to exceed the expertise by raw volume, and if they took the time and energy to curate it out the way they would need to to correct that they wouldn't be left with a large enough sample to actually scale off of.
Basically, either I'm dramatically misunderstanding something or the best we can hope for is the Average Joe on Reddit, who may not be a complete dumbass but definitely isn't a team of PhDs.