It's strange to think we chose to hunt or raise large animals; and to perform all that such a choice implies i.e. growing plants to feed them and more generally farming, when we could just raise ants and plants.
> in Gemini I even have a preamble that basically says "don't be a sycophant". It still doesn't always work.
Using this kind of strategy eventually leads to the LLM recurrently advertising what it just produced as «straight to the point, no fluff, no bullshit». («Here is the blunt truth»).
Of course no matter how the LLM advertise its production, it is too often non devoid of sycophancy.
Sometimes I draw UML-like diagrams when I join a project (and when the project is big enough in such a way my mind melts if I try to keep track of everything), I wonder if there are equivalent representations of such things.
Linear text is perfect to me for documentation, teaching/learning etc...
But also, systems seems to be better digested under the shape of spatial representations (I met a lot of CS persons that fantasized over the possibility of displaying all the files of your codebase in a VR-like environment augmented with visual cues (UML) and I must admit that I would find this unnecessary but comfortable -- and I can imagine applications of this in other domains; imagine a teacher arranging the whole of Kant philosophy as a spatial-visual system referencing positions, comments, etc..). Eyes are cool because you can focus on something while knowing that some available information is there around the zone you are focusing; in a sense, so is the hand, locally, but I imagine (I dont know) it would require some super-human level of braille reading to be able to jump back and forth reading on different fingers, so that's again a probably stupid question to ask to the blind crowd of hn : are you able to do this?
Businesses do business; but there are endavours to make tests be reliability indicators and in some (critical) domains you do write them to perform such a thing. write tests the way test-theory intended; as formal verification.
There is software for which writing code is a design act, and there is software for which you write specs before anything. I don't know if a) they are the same, b) they are different, c) one is better than the other.
As I am involved in more low level stuff, I prefer to read the source than the man pages, and I am very happy with people overcommenting their code as a user of e.g. a lib. On the other hand, it is unbearable to me to see comments on a codebase I am working on. Fortunately, emacs show/hide comments exists, so I find myself overcommenting things.
I think I fear system instability and its consequent (probable) problems.
Sleep does a lot of things, a lot I don't necessarily understand, all the OS layers are stressed at once, and with them a whole lot of other things, bith software and hardware related. Are all the drivers of your system trustworthy, are all the running applications trustworthy ? Are we sure no data loss will occur ? Will you lose audio, wifi, display or excessive battery because of a race condition or an error of some sort in an element of the whole involved stack ?
Most of the time everything is fine, or is it ? Maybe your computer will hit kernel panic after two hours or so, and you will have hard time figuring the real cause and origin.
tldr; I think it scares me because it increase the probability for the system to surprise crash at a very crucial time (while compiling something, in between two saves of a text buffer, during a write to disk...)
Seriously for my part, LLMs incarns exactly the only type of person that can break my nerves. Far too often I spot an hallucination, some bullshit rambling, sycophancy, or ----hughhhhh----- rethorical elements of language that makes me go mad :(.
examples for ---hughhhhh--- inducing stuff :
"I'll be blunt !"
"Here's the ground truth, no bullshit"
"Bottom line : <UPPER CASE EXPRESSION>"
"No fluff, technical, precise, no bullshit, devoid of unnecessary rethorical shapes, <etc..."