@drvolts@JoyceCarolOates Hey David, long time listener of volts here. Could you help me understand how that works? Most people I follow explicitly don't use the "for you" tab on here, myself included. I thought I was the sole curator of the things I see. How does it externally shape my world view?
A letter I wrote some months back for my 11-year-old daughter on what taste is, and why making stuff by hand still matters when AI creates such polished results. https://t.co/cYIcSg2oxS
@JFPuget She's the founder of PauseAI US. I suppose she has a very specific type of company in mind here: the ones racing to AGI/super intelligence. AI "labs" if you prefer
@everytstudies Not quite the same but have you had tangelos? I think the characteristic knob on top is because of its proximity to lemons. They've started to appear recently where I live (France). Interesting taste profile. https://t.co/qEd5LuHQNe
@princertitude Ça ne change pas l'argument de fond mais: la chaleur évacuée, elle vient de l'extérieur... Au premier ordre, il n'y a que la puissance consommée par le compresseur qui est ajoutée au "système ville". Sauf erreur, c'est encore un facteur ~4 supplémentaire en défaveur des véhicules
@pimarty86 Gemini 2.5 Pro en mode deep research. C'est extrêmement bien sourcé, et je ne l'ai encore jamais vu inventer une ref (bien que ce soit sûrement possible)
@eshear I think low-dose antipsychotics might be an answer. Information is scarce but it has a sort of GLP-1 vibe, as a repurposed class of drugs with surprising benefits. See quetiapine, olanzapine, paliperidone... Some seem to greatly improve sleep without diurnal side effects
@robinhanson The best (albeit not concise) articulation I've read comes from @everytstudies. E.g. https://t.co/AlaU18RiQX. But in a nutshell: STEM culture seeks to isolate phenomena to make sense of them, H instead tends to contextualize them
@adad8m That sounds like textbook confounding...? If you can afford to travel 4x / year, chances are you're wealthy, no? Depends on the exact statistic I guess
Another nice application of ML to analyse experimental combustion data. I'm grateful for this nice collaboration with colleagues from @KAUST_News! Paper: https://t.co/Lk1CCsBoMF
A long time in the making, new paper on re-imagining numerical discretization schemes on unstructured grids with deep learning. https://t.co/0XBlkY8OFv. Details in Luciano Drozda's PhD defense next Tuesday. https://t.co/G46EhLHqfy
New Paper is out! "Data-driven wall modeling for turbulent separated flows". https://t.co/waZPG9chsP Actually a prequel to a paper that came out last month (the magic of peer review timelines :) ): https://t.co/H0xav8FTvw
@NymMartin Naïvement je dirais que dans un labo on ne "fait pas de la science", on fait de la recherche. C'est pas un jeu de mots: la recherche ça couvre tout ce qui vise à un jour peut-être contribuer à la science, de l'étude du monde aux méthodes qui le permettent.