davebren

6 ms·

31.
by davebren 5mo ago
When you think about the word apple and what it signifies, what do you experience? Is there a feeling of "appleness"? Do you think that sense of meaning is equivalent to the numerical weights of an LLM?
32.
by davebren 5mo ago
Well yeah there is likely an equivalence between computability and epistemology, but I'm not sure it matters when comparing LLM intelligence to human intelligence. There is clearly a missing link that prevents the LLM from reaching bey
33.
by davebren 5mo ago
> What is the difference between what LLM's do and "true" understanding? The thing where you can understand the meaning of this sentence without first compiling a statistical representation of a 10 trillion line corpus of
34.
by davebren 5mo ago
> What is preventing AI from continuing to improve until it is absolutely better than humans at any mental task? No matter how much compute time it's given to combine training samples with each other and run through a validation eng
35.
by davebren 5mo ago
> I think that's just a matter of having them able to work on longer and longer time horizons. No this will never do the kind of math that humans did when coming up with complex numbers, or hell just regular numbers ex nihilo. No m
36.
by davebren 5mo ago
Pattern matching against millions of IQ test questions from a training set in order to score 150 on an IQ test doesn't give you an intelligence equivalent to 150.
37.
by davebren 5mo ago
Programming is the ultimate exercise in technical communication, even if we remove the aspect of communicating about the software to other programmers or non-programmers. It's not just communicating to the machine. The architecture, or
38.
by davebren 5mo ago
I'm not talking about the edge cases where it goes off the rails, I'm talking about the way it normally conversates. The way it was trained to do so through RL.
39.
by davebren 5mo ago
Please stop pushing the "getting left behind" fear. Prompting is not a skill, there is nowhere it's going that it can leave anyone behind. The "skill" of prompting is 100% dependent on one's skill programming,
40.
by davebren 5mo ago
> was able to vibe-code an MVP of our product that we've been working on in a matter of hours. > Given a little more time, I could solo ship and maintain what has taken us a few years to build as a small dev team. Sorry but this
41.
by davebren 5mo ago
I'm been in a community that makes a lot of cognitive training software. There's some core open source projects that were created without LLMs, but new projects are now mostly created by young people vibe-coding from scratch or fo
42.
by davebren 5mo ago
They're in a tough spot. They can train out the pretending to be human, sycophantic, lying, all-knowing aspects of the model, but this is how they got all the investors and CEOs on board the hype train. Psychosis is the product.
43.
by davebren 5mo ago
Mapping and photographing every road on the planet? Easy. Not manipulating our chatbot users into psychosis and suicide or worse? No way can't be done.
44.
by davebren 5mo ago
That was all based on the assumption that scaling LLMs would lead to AGI. That didn't happen and won't, it can be proven for anyone capable of digging into the details of what models actually are and what they are doing (I recomme
45.
by davebren 5mo ago
The loop on getting slop out to market quick in order to get feedback is already flawed. If you don't understand the problems of your customers well enough to come up with a coherent vision for how to solve them you shouldn't be t
46.
by davebren 5mo ago
If you want to declare yourself an NPC language model then go ahead but let the rest of us that experience meaning carry on.
47.
by davebren 5mo ago
Ah that looks very interesting I'll check it out, thanks.
48.
by davebren 5mo ago
Yep humans and civilization are subject to the same model-collapse phenomenon as they interact more with LLMs, but engineering knowledge has always been held by a small minority with certain personality characteristics. Maybe the minority w
49.
by davebren 5mo ago
> What exactly is so fundamentally difficult in 2026 that someone with years of experience can’t learn? It's having the experience to point out all the issues in the generated code that's the issue. The interns they hired won&#
50.
by davebren 5mo ago
That shouldn't apply to tech where there's generally always more market to capture and competitors looking to offer a better product and take your market share.
51.
by davebren 5mo ago
Not everywhere, and this is the year where the price per token will go up closer to the supply cost.
52.
by davebren 5mo ago
No, and nobody is going to find it when the guy from a third world village using a fake name that they hired to vibecode features puts a bitcoin miner in one of his daily 10k line agentic PRs.
53.
by davebren 5mo ago
By judging employees by how many tokens they burn.
54.
by davebren 5mo ago
It's more that the young people are more likely to go along with the hype because they aren't experienced enough to know the limitations of LLMs (their baseline skill level is too low to compare), they get to feel like they are mo
55.
by davebren 5mo ago
Young people today have grown up in a low trust society and have a totally different mindset. They have no qualms with figuring out how to extract as much as possible from their employer while providing as little value as they can. They wil
56.
by davebren 5mo ago
I know there's an unspoken rule to not speak directly about whatever shady thing your organization and leadership is doing but for this situation it would be surprising if there isn't some correspondence that would come up in disc
57.
by davebren 5mo ago
Yeah but we can't let the insufferable dictate our way of speaking. In spoken language I hear it mainly by people that don't have a scientific background trying to sound more scientific.
58.
by davebren 5mo ago
It actually is important and if I was hiring you I'd find it useful to get a more comprehensive understanding of your experience, especially if there's something I'm aware is a very challenging problem to solve. And it would
59.
by davebren 5mo ago
Probably gonna get downvoted for this, but when you give an anecdote you don't have to preface it with "anecdata, n=1 sample size". We know it's from your individual experience because it's a story about your indivi
60.
by davebren 5mo ago
Nah pretty much everyone picks red if it was a real scenario. It's also bizarre to call everyone picking red a sociopath because they don't want to gamble their life on some naive and idealistic view of humans.

More ›