First agreed. Honestly I haven’t an LLM do anything close to an engineer. You can get lucky for a code snippet or two and maybe a few ADRs but even those are fraught with the chance of slop clean up work. I am not arguing that.
The node is only probalistic because of the random number added. After the fact the number is known.
You can try to reason about it which is an arguable improvement from the former engineer you can’t get ahold of anymore. So you can could see that it triggered a node associated with c++ when it generated this string of tokens or compare it other generations to map it to a type of call.
Where it doesn’t compare to a human engineer at the moment is quality or novelty.
What I mean is for post hoc analysis since you know the coin flips and the inputs you can step through the layers and nodes calculation deterministically.
Neural nets are dumber than you thought for sure. What I mean is the coin flip results are known after the fact.
I mean probabilities in matrices are nodes in a neural net, right?
The only thing that makes it non-deterministic is the tempature value which is known after the fact from my understanding, so that should be able to deterministic.
Most of the LLMs architectures that I know of are neural net based.
Tbf the exact input and node activation is all you’d need for forensics, the only reason to refeed would old input plus new input mix, which should be new output.
TIL there are some unionized workplaces in SUSE that’s great!
I hate to say this but can’t LLMs solve the cognitive debt problem better then people? Wouldn’t having a stochastic machine that can have its state frozen in time, retrived at convience, and fed the same inputs be a potentially more transparent machine, thinking or otherwise?
I agree with you on stating thr problem though, and the term seems concise enough to me. The gap seems to be not in ability to do this but that creating graspable and reasonable audits for LLM usage is like many risk mitigation systems, an after thought in the industry.
Honestly. Been doing brief 1 hour no tech time at home now and it’s crazy how many times my girlfriend and I will do something, finish a conversation, etc, and then just go “well fuck” because we both sat at a computer or in a phone usage spot out of habit.
Context collapse and hallucination in the face of anything novel has been my experience 90% of the time. I even tried context compression by trying to do the research, test spec, plan, code tests, product spec, plan, code product method, but then the issue was the plan satisfied the AI too quick (“actually we already test all of the code, you can see that by reading the “we test all of the code” spec”).
I’ve been pro US imperialism for a while now, but honestly I’m glad that the rest of the world is shedding the worst of our power projection.
On one hand the heel position has really pushed a lot of the world in a positive direction, but there’s gotta be a better fucking way then, like you said destroying half a century of international relations and cooperation. That’s not to mention the bigger issue of US supported genocide in Palistein and the ethnic cleansing operation in the states, which I can’t excuse for any 4d accidently good things on the geopolitical scale.
Commercial support or if it’s critical enough infrastructure have a full time department/team to do it. Commercial support is the obvious choice for most countries starting out tbh.
Honestly SuSE, Ubuntu, and Hetzner all have good for enterprise support and its crazy that they don’t have more major gov business in the EU now.
But honestly yeah, pay for labour to maintain things. It’s not a breaking new idea. You can do that AND own the things instead of renting from private companies too. Also not a new idea.
The only thing new is that we are talking about computers and we have museums for them already so maybe get with the times, right?
The all in ones tend to be heat pump condensor dryers. Which are great, way less power used, and genteler on clothes, BUT take way longer.
I got one, and I struggle to think I’d ever go back to moving laundry over between machines unless I had to, even if it means 2-3 hours for medium load of laundry. It does that in the background anyways.
I kind of want to buy one know and start sending invoices for storage of their equipment.
This is common misunderstanding most people assume a climate pledge to protect it, they pledged to destroy. /s
I think when OPM was hacked that is what employees got lol
Oh true! It’s funny talking some data sci folks that are just “I didnt need GPUs but now I might as well optimize for them now” because they’re org has tons of that compute now for “AI”
AI development has a few factions. People that want to make and fuck their AI daughter/girlfriend. Researchers that want to under the basis of the human mind. Capitalists that want to convert human intelligence into capital they can fully own (slavery is the first attempt). Then finally cultists that want to make God in their image.
It’s a crazy field for someone that is fascinated by thinking about thinking and want to make useful little tools.
It’s not really coin flips it and random adjustments of weighted values that decides which nodes are activated which, plus random adjustments, influences the next node activation until out