AStrangeMorrow

5 ms·

31.
by AStrangeMorrow 7mo ago
Prompt quality does matter, but at some point context side does matter. I’ve had thing like a system that has a collection of procedural systems. I would say “replace the following set of defaults that are passed all around for system X (li
32.
by AStrangeMorrow 7mo ago
Yeah absolutely, at this point I also start new chats after 3-4 prompts. Especially with thinking models that produce so many tokens. Usually things go smoothly but sometimes I have situations like: “please add feature X, needs to have ABCD
33.
by AStrangeMorrow 7mo ago
I know, I find myself in this silly situation where I have to adjust my writing style because I write like an AI: always loved my bullet points and dashes. At work I also always tended to send slightly longer but structured answers. I found
34.
by AStrangeMorrow 7mo ago
Yeah. I feel like that like many projects the last 20% take 80% of time, and imho we are not in the last 20% Sure LLMs are getting better and better, and at least for me more and more useful, and more and more correct. Arguably better than
35.
by AStrangeMorrow 8mo ago
Yeah I helped build a 3D parametric engine a few years back at my company. This kind of stacked game of life was one of the first script I added to test it, alongside the 3D terrain generation using Perlin noise. I highly doubt I was the fi
36.
by AStrangeMorrow 8mo ago
Same here. As much as I enjoy a lot of the technical stuff, I never click on 80% if it because it is often “thing that already exist but in lisp/rust/etc, new tool similar to X to free/one extra feature/lightweight”. So
37.
by AStrangeMorrow 8mo ago
I disagree quite a bit. For me the medium, the technique, the process is all part of the art. Yet I still think the end result is also critical. But coming up with create ways to produce art matters. And I am confused about the “doing it an
38.
by AStrangeMorrow 8mo ago
Yeah it is easy to imagine the past as a set of very disconnected cultures when they were not. I remember hearing about people like https://en.wikipedia.org/wiki/Giovanni_da_Pian_del_Carpine He was a medieval Italian d
39.
by AStrangeMorrow 8mo ago
Yeah for having tried a few times, they only relative successes I had was having some World engine to manage the structure and style (generates character names and relationships, places names and location, biomes, objects, tracking world st
40.
by AStrangeMorrow 10mo ago
Yeah I am not an absolute GenAI hater. I’ve used it quite a bit myself, and I think there are ways to be creative about it. However 95% of what we see online, especially in the Ads space, is bottom of the barrel quality. It is obvious basic
41.
by AStrangeMorrow 10mo ago
Yeah I like NeuralViz too. Honestly the writing is what makes it funny in the first place. The AI imagery just adds that extra weirdness on top. Like my faves being the street interviews an the gluron that reviews the Zillow page etc
42.
by AStrangeMorrow 10mo ago
Yeah for me the three main issues are: - overly defensive programming. In python that means try except everywhere without catching specific exceptions, hasattr checks, when replacing an approach by a new one adding a whole “backward compati
43.
by AStrangeMorrow 10mo ago
It seems like it is fiction, from what I could find. I was doubting it at times but it feels like for how old it was some of the tech wasn’t quite there then. https://news.ycombinator.com/item?id=37723862
44.
by AStrangeMorrow 10mo ago
Yeah love her stuff. And honestly the voice description is part of the music flow at this point. I feel like that’s kinda how people imagined navigating whatever cyber domain when the first big cyberpunk novels came out
45.
by AStrangeMorrow 10mo ago
Obviously much simpler Neural Nets, but we did have some models in my domain whose role was to speed up design evaluation. Eg you want to find a really good design. Designs are fairly easy to generate, but expensive to evaluate and score. U
46.
by AStrangeMorrow 10mo ago
Yes, Levenshtein in that case give too big an exploration space. A keyboard edit distance would probably work better. Delete and swap are still 1 but replace and add should be within say 1-key at most
47.
by AStrangeMorrow 10mo ago
My guess is also that not all typos are equal. Should have a stricter edit version for 1-keystroke-away filtered edits (that is delete, swap or add 1 key away / replace one key away) instead of pure Levenshtein. Like Fqcebook is a more
48.
by AStrangeMorrow 10mo ago
If I understand correctly from the paper what qualifies as an edit distance of 1 is pure Levenshtein distance-1 right? Just curious because while the edit-1 space can be fairly big, I’d assume all edits have very different probabilities. So
49.
by AStrangeMorrow 10mo ago
Yeah, same as a French speaker first living in the US, I have to sometimes refrain myself from calling things “just fine”, “will do” or “not bad”. These are still used in American English, but I tend to use them for cases were people normal
50.
by AStrangeMorrow 10mo ago
I really enjoy writing some of the code. But some is a pain. Never have fun when the HQ team asks for API changes for the 5th time this month. Or for that matter writing the 2000 lines of input and output data validation in the first place
51.
by AStrangeMorrow 10mo ago
Yeah at this point I basically have to dictate all implementation details: do this, but do it this specific way, handle xyz edge cases by doing that, plug the thing in here using that API. Basically that expands 10 lines into 100-200 lines
52.
by AStrangeMorrow 10mo ago
Yeah some areas I looked at seemed fine (US, EU, Shangahi), but some the buildings are on top of the roads, and they look more like Perlin Noise than actual buildings. E.g main parts of Beijing. OpenStreetMap and such have pretty good build
53.
by AStrangeMorrow 11mo ago
I feel like to some degree, things have gotten less affordable. And I have seen a big push of the idea that “a job is just making money, find your happiness somewhere else”. Which led to more and more people looking for a job that pays well
54.
by AStrangeMorrow 11mo ago
Yeah people love to hate on Tesla and always point out the maintenance issues. But from the reports I have seen, Tesla is indeed one of the companies with the worst record in terms of number of incidents requiring maintenance, but among the
55.
by AStrangeMorrow 1y ago
I think, from my experience, what they mean is tool use is as good as your model capability to stick to a given answer template/grammar. For example if it does tool calling using a JSON format it needs to stick to that format, not hall
56.
by AStrangeMorrow 1y ago
Yeah even LLM generated code for a 2D optimization problems with many spatial relationships has being absolutely terrible, while I had great success in other domains.
57.
by AStrangeMorrow 1y ago
Yeah funnily I did a project where we had an LLM based interface to in house 3D parametric modeling system and it did fairly well. However I’ve been trying to use LLM, both as orchestrators and in other cases to write code for 2D optimizati
58.
by AStrangeMorrow 1y ago
Yeah I really liked that book, but honestly almost dropped it after the first chapter for that very reason. I get the point that is being made with all this gratuitous violence in the context of the whole book, but still
59.
by AStrangeMorrow 2y ago
Yeah sadly I have a similar experience from my school. People going into consultancy were either people that just did software engineering to quickly move into a management position and make more money, or people that were not passionate ab
60.
by AStrangeMorrow 2y ago
Yeah. Also might be a bias from my school, but for the software engineering specialty, in average the worse a student was performing the more likely they’d end up working for a consulting company after graduation. A few exceptions (student

More ›