a-dub

9 ms·

151.
by a-dub 9mo ago
my guess is that they are probably drowning in traffic since claude code really took off over the break and are now doing everything they can to reduce traffic and keep things up.
152.
by a-dub 9mo ago
yeah i was just wondering that. i wonder how much stem material is in the training set...
153.
by a-dub 9mo ago
flipping the signs on the logits would seem to give the "least likely" but i think in practice you're more likely to be just operating in noise. i would expect that tons of low probability logits would have tiny bits of energ
154.
by a-dub 9mo ago
it's really good these days. nvidia finally fixed their drivers (i suppose all it took was becoming the richest company on earth), kde is really nicely polished and all the friction from the x11 to wayland transition is over (at least
155.
by a-dub 9mo ago
the classic was dragon's lair: https://en.wikipedia.org/wiki/Dragon%27s_Lair_(1983_video_ga... laserdiscs were a weird 70s/80s analog optical video disc technology where many players had a db-9 (edit: just lo
156.
by a-dub 9mo ago
are you a completely autonomous agent?
157.
by a-dub 9mo ago
> Until I was asked, the 'seams' between my original logic and the AI’s linguistic polish were invisible. no they were not. to me it was obvious and that is why i "asked." this gets at a sort of fundamental misconcept
158.
by a-dub 9mo ago
seems like the real places to make durable changes would be in the desktop environment software packages (kde, gnome, etc)? although it also seems useful to figure out which third party packages and default settings make sense.
159.
by a-dub 9mo ago
did you use ai to write this as well?
160.
by a-dub 9mo ago
it always felt to me like the major innovation behind darksky was the clever exploitation of newly broadly available precision gps. prior to darksky, i'd look at the radar picture on my smartphone and the pin where i was to try and fig
161.
by a-dub 9mo ago
maybe it would work if they could encourage end users to be rigorous? (ie, detect if they have the capability to rate well and then reward them when they do by comparing them against other highly rated raters of the same phenotype)
162.
by a-dub 9mo ago
i haven't seen any rust projects yet that maximize embeddability, portability, space and transparency. maybe an argument could be made for the linux kernel stuff, but nothing like a sqlite that gets compiled and included in a million t
163.
by a-dub 9mo ago
i never understood why people like vscode. visual studio used to be great back in the day because it was really thoughtfully designed and everything was nicely integrated. vscode always felt like a hodgepodge of random features and extensio
164.
by a-dub 10mo ago
never played with windows 2 but looks pretty awesome. i think in this era the home market was largely saturated by home machines (atari st, amiga, apple 2+, a little macintosh). i don't remember a lot of pc juniors or other machines r
165.
by a-dub 10mo ago
> The actual code powering Gmail probably dates back to the late 80s or early 90s and has had several hundred thousands of hours of work put into it. no. google did not exist until the late 90s. various forms of internet email sure did,
166.
by a-dub 10mo ago
julia is still clunky for these purposes! you can't even plot two things at the same time without it being weird and there's still a ton of textual noise when expressing linear algebra in it. (in fact, i'd argue the type syst
167.
by a-dub 10mo ago
what can you do with bash that you cannot do with the python standard library and shelling out?
168.
by a-dub 10mo ago
maybe ask it to use 1990s table based layout approaches?
169.
by a-dub 10mo ago
i like to put it in live mode and point it at my plants and have conversations about how they're doing. it properly identifies them and flags any signs of disease and then provides correct next steps.
170.
by a-dub 10mo ago
as i understand: numerical methods -> smooth out noise from sampling/floating point error/etc for methods that are analytically inspired that are computationally efficient where monte carlo -> computationally expensive brute
171.
by a-dub 10mo ago
github actions is pretty terrible. the self hosted runner host is some horrific dotnet csharp mono monstrosity and "language" is some javascript wrapper nonsense that needlessly creates a half baked dsl around running basic shell
172.
by a-dub 10mo ago
i wonder how resistant it is to fine tuning that runs counter to the principles defined therein....
173.
by a-dub 10mo ago
they acquihired the team and derisked their investment in building claude code on top of bun. makes sense to me. moreover, now they can make investments in order to make it an an even more efficient and secure runtime for model workspaces.
174.
by a-dub 10mo ago
> However, there is an unexpected upside to having the menubar there even so. Because macOS apps can't not have a menu bar, they are forced to expose their commands there. Which usually ends up being a more stable UX compared to all
175.
by a-dub 10mo ago
there's some ubuntu/gnome thing that replicates the worst features of the mac. but here's the real question: why? the global menu bar is literally the most dated and outmoded element in macos. it isn't 1993 anymore. your
176.
by a-dub 10mo ago
macos feels dated these days to me. it's time for a clean slate.
177.
by a-dub 10mo ago
idk. i don't think there's a good story if you're primarily interested in full-cycle e-waste minimization and avoidance. the mac hardware is good and is designed to live on timescales like that (probably because they are prim
178.
by a-dub 10mo ago
could try asahi? i think it's pretty good on the m1. lenovos remain good if you get a high spec thinkpad. maybe get a few year old high spec thinkpad new/refurb off ebay with a three year service contract (search "p1 gen 6&q
179.
by a-dub 11mo ago
freebsd didn't have the hardware support base that linux did and suffered a huge delay in rearchitecture when x86 smp hardware became widely available. (only one cpu could be in the kernel at a time, the "bkl", was a major i
180.
by a-dub 11mo ago
i think in order to make this kind of argument you would need to be able to show all of the trajectories that are effectively reachable as a result of pre-training, and then how much effective pruning takes place as a result of total adjust

More ›