461–469 of 469 posts
For me, the biggest shift is people who don't care about local AI. The idea that you can no longer code without paying a tax to one of the billion $ backed company isn't sitting well.
Local could still be useful for chat and data processing though
If coding truly becomes effortless to produce - and by that extension a product becomes near free to produce - then I find it quite odd that the executive class thinks their businesses won’t be completely up ended by a raging sea of competition.
Not unlike all the A.I. companies all determined to build the machine god while predicting it’ll be disastrous. Same thing - better it starts with us
Earlier quoted context omitted.
You can enjoy doing woodworking without power tools but that's irrelevant to a job where people want it done fast with power tools.
Except a power tool works as it should each and every time. 100% reproducible. That's what's so great about it. Can we stop with the lazy analogies? Everyone's read some variant of this on here by now. Come up with something that's genius to read.
Earlier quoted context omitted.
You can enjoy doing woodworking without power tools but that's irrelevant to a job where people want it done fast with power tools.
Woodworking analogy for AI is not "power tools vs handsaw", its "power tools vs. wood 3D printer". You don't do any of the creating, you only ideate and allow the machine to do all the creating. It's simply not wood working anymore. Its something else entirely
Earlier quoted context omitted.
The marginal differences in quality seem pretty meaningful right now, enough to make Claude wildly dominant, but some of the locally runnable models like Qwen feel only a few months behind the leaders. I'm betting the generational gains level off and smaller local models close the gap somewhat. Then harnesses will generally be more important than model, and proprietary harnesses will not offer much more than optimiza…
Qwen3.5 really is great. But the biggest thing is going to be context. Whilst a 10gb card can run a 9b model with some context .. for coding you really want a lot of context. So if paying 200 a year for 1T in context, vs your 32k context.. that's the thing I see as being the driver. Personally ive found great success with using open code, having Opus as my plan agent, and omnicoder-9b as my build agent. Get opus to p…
After having gone all-in on LLM agents for a while, I'm not so sure anymore. An LLM with lots of context can sometimes generate more accurate code, but it can also hide decision-making from you, the person who actually has to maintain that code. If the LLM pulls in 1000 files to make a decision, that's no longer a decision that you can understand.
Earlier quoted context omitted.
I am building them because I am using them to do my work faster. I'm not selling anything, but I can see the quality of what is created and it is on-par with much of the stuff on the App store. No one would even notice that it is a co-creation unless I mentioned the time to create it. Just to be clear. Vibe coding implies that you are not reviewing the code that is created, or even knowing what is being created. That…
> I am building them because I am using them to do my work faster. You sound like CTO at my company rewriting stable libraries in languages he is familiar with and calling it 100x...
Earlier quoted context omitted.
My impression is that people who think that LLMs will completely release reviewing or writing code have never really worked on anything safety critical. I'm not looking forward to the next wave of pacemaker glitches.
You act like we live in a world where companies are held sufficiently liable.
Earlier quoted context omitted.
You're anthropomorphizing. LLMs can't lie nor can they tell the truth. These concepts just don't apply to them. They also cannot tell you what they were "thinking" when they wrote a piece of code. If you "ask" them what they were thinking, you just get a plausible response, not the "intention" that may or may not have existed in some abstract form in some layer when the system selected tokens*. That information is go…
>You're anthropomorphizing. That is fine. You should, and you'll get the best results doing so. >LLMs can't lie nor can they tell the truth. These concepts just don't apply to them Nobody really knows exactly what concepts do and don't apply to them. We simply don't have a great enough understanding of the internal procedures of a trained model. Ultimately this is all irrelevant. There are multiple indications that t…
There are however limitations imposed by the architecture. An LLM cannot form secret chains of thought (though in theory a closed system outside the end-users' control could hide tokens from at least the user), nor can it model decent metacognition. They also have an at-best weak concept of fact vs fiction in general, which is why we get hallucinations. All of that isn't exactly optimal prerequisites for telling lies.
Also your car isn't a coward because it refuses to run into an obstacle onboard systems detect. The car's designers may have been cowards. Your car also isn't a hero for protecting you during a crash. Neither are LLMs virtuous or liars. If some AI company went out of their way to intentionally construct an LLM such that it outputs untruths, it's not the LLM that is lying to you, it's Open AI/Anthropic/whoever you're interacting with. You're using their system. They are responsible for what it does. If it tells untruths they may have automated the act of telling lies, but it's still them doing it.
> There are multiple indications that the same can be said for humanity, that we perform actions and then rationalize them away even without realizing it
I was hoping to get a response like yours, because I'm genuinely curious about where it leads.
I believe what you said is true in the general sense, where we solve easy problems subconsciously in parts of our brains dedicated to supporting the conscious mind, without then being able to explain how we did it.
However this is a lot less true for engineering tasks, which have a lot more active planning. Sometimes software development means just being a fancy constraint solver, finding a solution that works while applying some best practices. When pressed why one chose that particular solution, one might be tempted to post-hoc rationalize it as the best solution, even though it was just one that fit. But that's merely making it out more than it was, not taking away from the accomplishment of finding one that worked, which likely required some active thinking.
At the other end of the spectrum is making architectural decisions and thinking ahead as one creates something novel. I would be able to tell you why everything exists, especially if I merely added it in anticipation of something that will use it later. There's a ton of conscious planning that goes into these things.
Most coders are still turning over problems they're dealing with at work in their head when they're going to sleep late in the day. This is very much the opposite of solving problems subconsciously.
Earlier quoted context omitted.
>You're anthropomorphizing. That is fine. You should, and you'll get the best results doing so. >LLMs can't lie nor can they tell the truth. These concepts just don't apply to them Nobody really knows exactly what concepts do and don't apply to them. We simply don't have a great enough understanding of the internal procedures of a trained model. Ultimately this is all irrelevant. There are multiple indications that t…
> We simply don't have a great enough understanding of the internal procedures of a trained model. There are however limitations imposed by the architecture. An LLM cannot form secret chains of thought (though in theory a closed system outside the end-users' control could hide tokens from at least the user), nor can it model decent metacognition. They also have an at-best weak concept of fact vs fiction in general, w…
This is not an architectural limitation. All the Frontier Labs have pretty much noted you actually have to work to keep RL-tuned thinking models with human readable chains of thoughts. They quickly devolve into (still working) gibberish pretty quickly otherwise. And there are a few research variants out there that keep chains of thought entirely internal. It's not a thing for frontier models because frontier labs do not want secret chains of thought, not because it's an architectural impossibility.
>I would be able to tell you why everything exists, especially if I merely added it in anticipation of something that will use it later. There's a ton of conscious planning that goes into these things.
I’m not denying deliberation. I’m denying that introspection gives you transparent access to the mechanism of deliberation. Those are different claims. You may indeed have spent hours consciously turning the problem over. But when you later say “this is why I did X,” that report may still be a compressed narrative laid over a messier process involving habit, salience, constraint satisfaction, tacit pattern matching, and only partly articulated foresight. The point is not subconscious versus conscious. The point is that reportability is not the same thing as causal access. And the evidence that our reportability is faithful is much weaker than people usually assume.