sailingparrot

5 ms·

61.
by sailingparrot 6mo ago
Gas Town really feels not just vibe coded but also vibe designed. I looked into it, to see whether multi agent setups really made a difference, the entire design philosophy feels like it was « let’s add one more layer of agent and surely th
62.
by sailingparrot 7mo ago
Warms my heart to learn their families will finally be able to afford nutritious meals, put clothes on their backs and maybe even afford a bike to go to school rather than walking 2h everyday. We need more uplifting stories like this one. T
63.
by sailingparrot 8mo ago
> This question IMO reveals how the abstraction of numbers can imprison our minds. Is it the abstraction of number that imprison our mind or just the reality of having a job and other social constraint based on all of us agreeing on a ti
64.
by sailingparrot 8mo ago
Yes, that’s what I meant and yes I was referring to the N line in the sunset as the worst offender (in my commutes)
65.
by sailingparrot 8mo ago
Sorry, in the case of the bus there are too many bus stops (although there are more express lines now), so the bus stops a lot instead of having less stops where more people get off and walk one more block (what the article talks about). Th
66.
by sailingparrot 8mo ago
> Two, removing stops will likely not make the remaining stops nicer. Cities aren’t thinking about how to allocate a fixed bus budget. But that’s not at all what the article is about? The thesis is not that having bus stops with music an
67.
by sailingparrot 8mo ago
> it builds production apps up to 4x faster and produces client bundles up to 57% smaller. I suppose that is you being overly retributive indeed.
68.
by sailingparrot 8mo ago
> let every part of the market slip away. Which part of the market has slept away, exactly ? Everything you wrote is supposition and extrapolation. Nvidia has a chokehold on the entire market. All other players still exist in the small p
69.
by sailingparrot 9mo ago
> How much maintenance do you need? A lot. As someone that has been responsible for trainings with up to 10K GPUs, things fail all the time. By all the time I don't mean every few weeks, I mean daily. From disk failings, to GPU over
70.
by sailingparrot 9mo ago
> Nvidia has been using its newfound liquid funds to train its own family of models Nvidia has always had its own family of models, it's nothing new and not something you should read too much into IMHO. They use those as template ot
71.
by sailingparrot 9mo ago
> you don't need to make a video model. You probably don't need to decode the latents at all. If you don't decode, how do you judge quality in a world where generative metrics are famously very hard and imprecise? How do y
72.
by sailingparrot 10mo ago
It was exciting because of what it means regarding how a model learns, regardless on whether or not its commercially applicable.
73.
by sailingparrot 10mo ago
Which words am I putting in your mouth? > HPV causes genital warts False. Not all do. And more importantly, the ones that cause cancer do not! > Once you are confirmed HPV positive (again, you won't be confirmed without getting g
74.
by sailingparrot 10mo ago
You are giving some honestly really bad and dangerous info. The HPV strains that cause cancer and the ones that cause genital warts are different. The strains that cause cancer do not cause warts. So you can very much have HPV without genit
75.
by sailingparrot 10mo ago
You must be looking at non tech positions, most of their research/applied role go up to ~550k, and they do offer more than advertised for strong candidates. + hiring cash bonus + equity (which is a lot). https://openai.com&
76.
by sailingparrot 10mo ago
Uncharitable take. His last public stance on this a few months ago when he released nanochat was that he didn’t use coding LLM for it, even though he tried, because they were not good enough and he was just losing time, so coded everything
77.
by sailingparrot 10mo ago
The author is talking about progress within one game/set/race not over the long term. In lifting during your first rep you can notice that your stance was slightly too wide and adjust it for the other 4 reps. Just like in rowing i
78.
by sailingparrot 10mo ago
Just for training and processing the existing context (pre fill phase). But when doing inference a token t has to be sampled before t+1 can so it’s still sequential
79.
by sailingparrot 10mo ago
Yes, very few new NPP have been built in Europe recently. Quite a few have been built by Europe however. The french company Framatome alone, with 18k employees, is actively building 2 EPR reactors in the UK (+ preliminary studies for 8 more
80.
by sailingparrot 10mo ago
> So you want to create a completely new industry. From the ground. With all existing experts having retired. This is an article about Europe. Do you really believe France alone is operating 57 nuclear reactors, and producing 70% of its
81.
by sailingparrot 11mo ago
Sure, I'ld rather be doomscrolling on tiktok than being stuck for 4 years in trenches in France during WW1, we are talking about larger trend across time though. > I just don't trust any measure that's over 20 years old.
82.
by sailingparrot 11mo ago
> I guarantee our hunter-gatherer ancestors felt all the same emotions—burnout, comparison, envy, anxiety, stress, overwhelm, hopelessness. Yes they felt all the same emotions. You absolutely cannot guarantee they felt them in the same p
83.
by sailingparrot 11mo ago
> Throughout time, people have complained that technology is ruining the world. Before AI it was the internet, and before that it was TV, nuclear power, and so on. What if throughout time they have been right ? Any proof thst while tech
84.
by sailingparrot 11mo ago
> So basically they are admitting that not enough people will pay for it to be a profitable business Since when capitalism is about stopping trying to make more money right when you become profitable? If they can find a way to make 10x t
85.
by sailingparrot 11mo ago
The unbearable pain of having to handle bills of different sizes, there is not enough empathy in this world to truly pay hommage to your suffering.
86.
by sailingparrot 1y ago
At no point have I argued that LLMs aren’t autoregressive, I am merely talking about LLMs ability to reason across time steps, so it seems we are talking past each other which won’t lead anywhere. And yes, LLM can be studied under the lens
87.
by sailingparrot 1y ago
> If it conveys the intended information then what's wrong with that? Well, the issue is precisely that it doesn’t convey any information. What is conveyed by that sentence, exactly ? What does reframing data curation as cognitive h
88.
by sailingparrot 1y ago
> entirely embedded in this sequence. Obviously wrong, as otherwise every model would predict exactly the same thing, it would not even be predicting anymore, simply decoding. The sequence is not enough to reproduce the exact output, you
89.
by sailingparrot 1y ago
train on bad data, get a bad model
90.
by sailingparrot 1y ago
Yes, it's "just" an optimization technique, in the sense that you could not have it and end up with the same result (given the same input sequence), just much slower. Conceptually what matters is not the kv-cache but the atte

More ›