hansvm

41 ms·

661.
by hansvm 1y ago
Busses are huge though. You can get fares cheaper than the cost of gasoline for an equivalent single-passenger trip just by using the available seats and ignoring standing room completely. The seats are pretty wide, so that satisfies the an
662.
by hansvm 1y ago
That's a fair point. I think maybe the issue is one of the reference point we're implicitly choosing for creativity. "{{{{}}}}" is creative relative to our expectations for the problem -- falling outside the usual distri
663.
by hansvm 1y ago
> definitions Sure, perhaps. Take it up with the authors. > make sense...goal That's not necessarily the goal. Alignment definitely filters the available response distribution, but the result of alignment and fine-tuning can be h
664.
by hansvm 1y ago
A weaker statement is that creativity is bounded by entropy. The LLM is still free to respond "Four," "four," "{{{{{}}}}}," "iv," "IV," etc. A sufficiently low-entropy response cannot be cre
665.
by hansvm 1y ago
https://github.com/axboe/liburing/discussions/1047
666.
by hansvm 1y ago
> defined as syntactic and semantic diversity
667.
by hansvm 1y ago
The last time I did anything like this, the easiest workflow I found was to use your favorite high-level runtime for training and just implement a serializer converting the model into source code for your target embedded system. Hand-code t
668.
by hansvm 1y ago
And states like Rhode Island trying to actually charge truckers for those damages keep having their efforts struck down in the courts.
669.
by hansvm 1y ago
To be somewhat fair, their numeracy scatter plot visually looks more like a dart board than any real process. The fact that you got a positive-sloping line out at all from the regression has to do more with the positions of the outliers tha
670.
by hansvm 2y ago
The "fix" I've seen them testing is changing the wording to "show less shorts."
671.
by hansvm 2y ago
> too many cases to analyze One of my early-career successes was just creating a framework for generating every permutation of perf optimizations for every (log-scaled -- clz is very fast) input size and checking which was best, dropping
672.
by hansvm 2y ago
Figuring out which of two things is bigger is hard. The article lightly touches on the two main reasons: - Sometimes the definition is complicated. E.g., I'd accept a computer 10x slower at most tasks if a particular small subset were
673.
by hansvm 2y ago
> why science in particular? Because we can't usually measure our goals directly. We want outcomes like relativity and the two-slit experiment. Those results take a lot of time to uncover and have a meaningful chance of failure. If
674.
by hansvm 2y ago
Due to it being electric. That's really interesting. What prevents shorts?
675.
by hansvm 2y ago
"Safe" is relative, but I've taken older Honda civics through water part-way up the doors. When you're in the middle of nowhere it's nice to have options. Do you run the risk of major electrical faults if you run th
676.
by hansvm 2y ago
Sorry I wasn't clear. You can get a motor with 100k miles from a totaled car for $3k, including the labor to replace it. To your actual question, I bought mine (2008, manual) in 2018 for $5k with 100k miles in The Bay, and it took abou
677.
by hansvm 2y ago
I missed that, thank you.
678.
by hansvm 2y ago
I definitely worry about the ability of this thing to drive through a couple feet of water safely.
679.
by hansvm 2y ago
The Honda Fit is great. You can probably squeeze an extra decade out if you're willing to swap out the motor or transmission (used, 100k miles or so, if you shop around $2k-$3k should be doable), and if you're using it heavily the
680.
by hansvm 2y ago
Isn't this sort of thing illegal as a new vehicle in US markets because of those backup camera laws?
681.
by hansvm 2y ago
How do VSCode refugees impact a project like Zig?
682.
by hansvm 2y ago
It has Germanic roots too. Are any of the people you were close to growing up from the Midwest or from sort of Germanic or Scandinavian background?
683.
by hansvm 2y ago
There's a tendency sometimes to leave certain words implicit. "It does not [even] take 21 days to form a habit." Some of those linguistic habits never leave their birthplace, giving you colloquialisms like "do you want t
684.
by hansvm 2y ago
I've tried a few times (whenever somebody sent that sort of gift card my way), and it was always >$10 in explicit fees, plus all the menu items being silently marked up 20%. Mind you, that's still not exactly "expensive&qu
685.
by hansvm 2y ago
Uber's been bitten by that sort of thing before, though last time [0] it was even more obviously their own fault. [0] https://www.reuters.com/legal/litigation/uber-loses-appeal-b...
686.
by hansvm 2y ago
We have laws about false advertising and such things. At a minimum, there's a case to be made with respect to the warranty of merchantability.
687.
by hansvm 2y ago
The normal use case for `inline for` is when you have to close over something only known at compile time (like when iterating over the fields of a struct), but when your behavior depends on runtime information (like conditionally assigning
688.
by hansvm 2y ago
Some related fun facts: 1. That roofline curve idea applies to multiple processes, computers, and data centers just as well. If you have enough "cache" (disk, RAM, whatever), you can do a distributed matmul and actually effectivel
689.
by hansvm 2y ago
In the nicest way possible, do you have anything to add other than just saying that I and the rest of the world are wrong and insinuating that I made a claim I believe to be false? There's a lot of uncertainty in the world, but the cor
690.
by hansvm 2y ago
You could, but the willingness to frame it that way at a local decision-making level causes problems for global optimization. It's perfectly fine for a surgeon to triage patients and actions. It's much less fine for a hospital adm

More ›