Paul Ganssle @pganssle@qoto.org

@PalPartaca Hace como año y medio probé varias de alternativas al Kindle Paperwhite y al final me pillé un Pocketbook Era, que me a gustado mucho. Pesa un poco más que los otros, pero me parece una plataforma mucho más abierta. Si eso te da igual, el Kobo Libra Color también me gustó suficientemente y pesa menos.

@bitecode Though also skipped tests wouldn't help with coverage metrics, and presumably would be detected by lack of coverage, no? It is a strong argument for including tests in coverage metrics: https://nedbatchelder.com/blog/201908/dont_omit_tests_from_coverage.html

xfail might have that problem though, which is why I exclude xfailing tests from coverage: https://blog.ganssle.io/articles/2021/12/xfail-coverage.html

@bitecode Amateur hour. Doesn't the AI know failing tests should be marked xfail? 🤦

@CleoQc TBH I may be pushing the limits of blender technology, but I do not want some watery fruit juice, I want a thick iced fruit slurry and I don't think that is too much to ask. Somehow the açai bowl places always manage to pull it off.

@CleoQc I put milk in there, but I've made them the same way with all my other blenders and my Blendtech blender lasted 10 years (... before it set itself on fire when I was making a smoothie).

@kevin @CleoQc Possibly? I don't know that I would particularly like it even if it worked properly, though.

@CleoQc Glad that yours works for you, though. 😀

@CleoQc Just first-hand experience. I got a Vitamix last month and it has not once successfully made a (frozen fruit) smoothie without stalling out, in addition to the normal problem most blenders have where the contents get all bunched up and the blades spin freely.

This is the one time the "just use UTC everywhere" people are right (for my time zone).

@hugovk Despite the fact that the thing preventing a 3+-only release at this point is just me writing documentation (and thus this makes me feel very guilty), I'm kind of OK with this as a general principle.

I usually tell people not to use the parser at all. Relativedelta has a bit great UI but is a pretty good basic component. Would love the time to make everything faster, but, uh, I haven't even had time to write down how to use the things I already did so.. yeah..

@glyph @nedbat @brianokken I am not sure but they feel slightly different to me. I think of a benchmark as a fixed task that you do to track progress. Evals could be stuff like having humans rate 1000 examples of output, or using a specially trained evaluator model to rate a bunch of outputs. Probably a wide enough definition of either term allows one to encompass the other, but they probably have different centers of mass.

Ok yeah that is actually pretty loud. This is the power spectrum with the tip held about an inch away from the microphone. Going "aah" at the same microphone from the same distance at a normal or slightly elevated speaking level is giving me peaks around -40 and this thing is doing -20.

I think it is this one, in case anyone wants to... uh... light candles 😉 https://www.amazon.com/VEHHE-Electric-Rechargeable-Flexible-Barbecue/dp/B0852VJ6RD