

At least the lone crypto bro is getting appropriately roasted. They’re capable of learning.
At least the lone crypto bro is getting appropriately roasted. They’re capable of learning.
As the bioware nerd I am it makes my heart glad to see the Towers of Hanoi doing their part in this fight. And it seems like the published paper undersells how significant this problem is for the promptfondlers’ preferred narratives. Given how simple it is to scale the problem complexity for these scenarios, it seems likely that there isn’t a viable scaling-based solution here. No matter how big you make the context windows and how many steps the system is able to process it’s going to get out scaled by simply increasing some Ns in the puzzle itself.
Diz and others with a better understanding of what’s actually under the hood have frequently referenced how bad Transformer models are at recursion and this seems like a pretty straightforward way to demonstrate that and one that I would expect to be pretty consistent.
That would be the best way to actively catch the cheating happening here, given that the training datasets remain confidential. But I also don’t know that it would be conclusive or convincing unless you could be certain that the problems in the private set were similar to the public set.
In any case either you’re doubledipping for credit in multiple places or you absolutely should get more credit for the scoop here.
The thing that galls me here even more than other slop is that there isn’t even some kind of horrible capitalist logic underneath it. Like, what value is this supposed to create? Replacing the leads written by actual editors, who work for free? You already have free labor doing a better job than this, why would you compromise the product for the opportunity to spend money on compute for these LLM not-even-actually-summaries? Pure brainrot.
New rule: you’re not allowed to tell people to shut up and look at the numbers unless you’re actually good at math.
As usual the tech media fails to consider the possibility that part of the reason for Anthropic poaching people with promises of more money and huffable farts is to get this exact headline to try and get another round of funding from the VCs.
🎶 We didn’t start the fire
We just tried to profit
From our own new market
We didn’t start the fire
Though I see why we might’ve
I did not ignite it 🎶
I also appreciate how many of the “transformative” actions are just “did a really good thing… with AI!”
HR reduced time-to-hire by 30%! How? They told Jerry to stop hand-copying each candidate’s resume (I sleep). Also we tried out an LLM for… something (Real shit).
Like, these are not examples of how AI adoption can benefit your organization and why being on board is important. They’re split between “things you can do to mitigate the flaws in AI” and “things that would be good if your organization could do” and an implication that the two are related.
A) “Why pay for ChatGPT when you could get a math grad student (or hell an undergrad for some of the basics) to do it for a couple of craft beers? If you find an applied math student they’d probably help out just for the joy of being acknowledged.” -My wife
B) I had not known about the cluster fuck of 2016, but I can’t believe it was easier for the entire scientific establishment to rename a gene than to get Microsoft to introduce an option to disable automatic date detection, a feature that has never been actually useful enough to justify the amount it messes things up. I mean, I can believe it, butI it’s definitely on the list of proofs that we are not in God’s chosen timeline.
Possibly OT, but fits in with the “finance ruins everything” motif we’ve got going here:
My wife and I have been playing Stardew Valley again, and now the algorithms occasionally find us things like this
The continuing presence of stories like this is making me reevaluate my assessment that GenAI will never be good enough to replace creatives, not by estimating that the tech will be better but by adjusting down the level of competency that is apparently permissible. Like, anyone in a vaguely creative sphere who wants to start phoning shit in as aggressively as possible should probably do it if they aren’t already.
That was great! Thank you for putting in the effort to write it up.
I guess that’s fair. I was focusing in on his attitude towards craft, which seems incompatible with actually taking pride in doing a good job as opposed to simply skating by. But while I still take issue with his attitude there and want to give him a clockwork orange-style refresher about tech debt I think a bigger problem is that he’s taking predictable problems of the median programmer trying to use these systems and saying, effectively, “get gud”. This is especially galling given that the tech here is going to replace or supplant the kind of junior developer roles that allowed fresh graduates to actually get that experience that allows you to shepherd the next generation of junior devs (or I guess LLM assistants now).
I mean he accurately predicted the kind of dystopian shit Peter Thiel would do with a morally indefensible amount of money, so that’s something.
That’s a whole lotta words to say “I’m a bad programmer who aspires to be a bad manager of a team of programmers.”
Not gonna lie, if Rian Johnson ends up being in that milieu I’ll be absolutely heartbroken.
I mean I don’t doubt that some folks on the internet were absolute bastards about it. At the same time, while I’ve got a lot of love for Rian Johnson’s work and don’t have any room to criticize the process that creates it, I do have concerns. First off, while it’s artistically satisfying and a good personal defense, retreating into a bubble away from criticism doesn’t stop the economic and social repercussions of that criticism, which can definitely reflect back on the artistic product as it did when the far less interesting JJ Abrams was brought back to do the last Star Wars movie instead of letting Rian keep going. I don’t have a good solution for that, since fighting the internet hate machine isn’t something I’d wish on anyone, but it’s still a problem. This is especially the case with Gen AI here because the economic and social consequences that technology has on artistic production and creativity are the whole point of the criticism. Like, it’s not just that AI art is bad - we’ve seen plenty of bad art from human beings make it to theaters. Even if it gets less bad it’s replacing actual people with artistic visions and actual lives with a machine that is, somehow, even more of an environmental disaster and economic drain on society. It sounds like this is the kind of story that might be trying to engage with some of that in a meaningful way, but I don’t think that justifies actually using it here. Like, if you’re paying to enter the torment nexus in order to post up your propogands about how we shouldn’t have created the torment nexus, you’re still paying the fuckers who created the torment nexus for their creation of the torment nexus.
Godspeed my friend. What was the turnaround, if I may ask? How long was the interview process etc?
If that’s true then how has he maintained whatever passes for his career in Sci-fi whining these days?
That’s probably true, but it also speaks to Ed Zitron’s latest piece about the rise of the Business Idiot. You can explain why Wikipedia disrupted previous encyclopedia providers in very specific terms: crowdsourced production to volunteer editors cuts costs massively and allows the product to be delivered free (which also increases the pool of possible editors and improves quality), and the strict* adherence to community standards and sourcing guidelines prevents the worse loss of truth and credibility that you may expect.
But there is no such story that I can find for how Wikipedia gets disrupted by Gen AI. At worst it becomes a tool in the editor’s belt, but the fundamental economics and structure just aren’t impacted. But if you’re a business idiot then you can’t actually explain it either way and so of course it seems plausible