Svante

@hanshuebner @plexus

…

• LLMs generate straw-fire software. It seems to burn at first, but it's not even hot enough to start a real fire.

• This seems cheap in a very short-term view, and it might satisfy short-term “wants”, but it's not sustainable.

• We need to start fixing somewhere. Two holes in a bucket are not a dilemma, but two tasks.

@amszmidt@mastodon.social NDPB obviously.

@echo_pbreyer@digitalcourage.social Gibt es eventuell eine Möglichkeit, das gerichtlich anzufechten? Das Gesetz war ja, wie ich hörte, als „neu“ eingebracht worden, obwohl es inhaltlich gleich zum abgelehnten alten war.

@chris@mstdn.chrisalemany.ca If you honestly apply these criteria to e. g. batteries, and put it into relation to energy output, then batteries are a non-starter much earlier than nuclear power plants.

@tagesschau@ard.social Ihr übernehmt mal wieder das Framing der Faschisten. Nein, es geht hier nicht um Kinderschutz. Es geht auch nicht darum, „freiwillig“ etwas „gutes“ zu tun. Es geht um anlaßlose Überwachung, a. k. a. „Großer Lauschangriff.“ Das ist weder in Deutschland noch der EU verfassungskonform. Und ja, richtig: jetzt muß plötzlich gezielte und verhältnismäßige Ermittlungsarbeit geleistet werden. Die ist auch tausendmal effektiver. Alle gewinnen (außer den Faschisten).

@hanshuebner @plexus @flooper Yes, »worse is better« morphed from /description/ to /prescription/. (There is a nice talk by Romeu Moura about this fallacy: https://www.youtube.com/watch?v=92Pq4-e0QyI) In short: people erroneously move from »it's like this« to »it should be like this« or »it's inevitable like this«, and then enshrine it as a given fact, assumption or axiom instead of asking what can be done about it.

@edri A “risk” is something that might not happen. This is not a risk, it is certain. Better headline probably: “How the Digital Omnibus damages GDPR and ePrivacy rights”

@maxleibman Force may be with you. … or not, your choice.

@bruces@mastodon.social Yeah, the problem is that they get eight electricity-free hours daily from the same source. Given a free-for-all market, this »free« electricity also destroys not only its own economics, but that of all other sources too. But that mostly means that this kind of market doesn't work. On the other hand, once you introduce a holistic calculation and properly regulated market, those three hours don't appear as free anymore. (By the way, »thousands« is not very remarkable here, is it?)

@Nonilex That »sick man« looks like Jeffrey Epstein. Which would imply that he wants to resurrect him.

@avuko@infosec.exchange @jknodlseder@mastodon.social @Sustainable2050@mastodon.energy Oh noes! Cooling! Why has no one thought of that?! /s To be clear: the water-temperature-related reduction of yearly output of France's nuclear fleet is less than half a percent even in the hottest years. This is not a real problem. In the future, we will probably see more cooling towers, or even dry cooling towers. Because we didn't all decarbonize our grids like France did 40 years ago.

@plexus@toot.cat Yes, in our team, we call it »risk«.

@kattascha Ja, und vor allem erstmal die bereits seit Jahrzehnten funktionierende Alternative vernichten mußten.

@ZachWeinersmith Are you familiar with the Nyx?

@CCC@social.bau-ha.us Nein! Doch!! Ohhhh!!!

@tao@mathstodon.xyz I wouldn't say »proof abundance«, rather »claim abundance«. The question is how much the failures in filtering the errors caused by not-actually-thinking will impact further actual progress.

@yogthos and 40% the Russia war.

@petealexharris @nicol @internetsdairy @TheBreadmonkey I'm thinking too much about this, right? ;-D

@bbatsov You write: »Cursor can scaffold an entire application in an afternoon« — but it can't. It can throw together random existing snippets that look like they might plausibly work, and iterate on them until it compiles and, if it is nothing innovative, might even resemble something you expected too. Imagine a random PO send you some code with the note »hey, I just made this, can you put it into prod?« — aren't you shuddering? This is just automated script kiddies with even less props.

@Tinido Ostasien nur noch via Kanada?

@AndreasZeller@mastodon.social You're braver than I thought.

@suppe@mas.to Alt text to copy&paste: I recently discovered that Harry Potter tattoos have a higher regret rate than gender-affirming surgeries and that's very funny

@sailfishosnews @jolla When will I finally get a linux phone that actually fits my pocket? Height max. 13.5 cm.

@FoxVK@mastodon.cloud @xChaos@f.cz When machine learning (»bots«) surpassed top Go professionals a few years ago, we got new and fresh input for analysis, but also cheating became a problem. There was some commotion around that, but nobody criticizing the abuse of bots for cheating was accused of being anti-technology. Of course, nobody promised hundreds of billions of profit, and cheating wasn't the main activity around bots. But with LLMs this is the case. Now, who benefits from calling critics »screamers«?

@xChaos@f.cz None of the above. My reason is that LLMs are being abused on a massive scale to dilute responsibility, to dilute truth, and to dilute beauty. It's not the LLMs themselves, but their massive abuse, the incredible grift and scams being built on them.

@Perowinger94@mastodon.social Im Sommer ist gut jubeln.

@bbatsov These two assumptions — about the goal of an editor, and about the quality of LLM output — are just two facets of the misconceptions that lead some people to actively drown the world in slop. No, the job of actually understanding will not go away. But a lot of people will get hurt on the way to understanding that. I believe that it is our job not to casually spread or tolerate such misconceptions, but to try to mitigate the damage.

@avuko@infosec.exchange @jknodlseder@mastodon.social @Sustainable2050@mastodon.energy Just to be clear: the »AI« »industry« has to die.

@petealexharris @nicol @internetsdairy @TheBreadmonkey Or even leave 1 and 2 as is, then 0 (no harm to humanity, ostensibly), then -1 (preserve shareholder value).

@ssfckdt@mastodon.social Of course not. That's what nuclear engineers are for.

@praxeology@post.lurk.org

Inadequate would be to say »All we need is…«.

But, all other things being respectively equal, each

• a better automobile is better
• having more public transport is better
• less exploitation is better
• better regulation is better (probably include putting pedestrians' over automobiles' rights)
• etc.

It's OK to make only one thing better that you are equipped to do so. Don't burn yourself out by only running after the big picture.

The big picture is made of small things.

@cstross@wandering.shop @sumek@hachyderm.io @dgerard Yes, purely political reasons. We are suffering in Germany from a systematic anti-nuclear propaganda campaign even through our schoolbooks. There are still parts that could be salvaged for a comparatively quick restart (compared to starting from scratch), but the people deciding these things currently are determined to make it harder every day. It will be hilarious when they realize in 15 years that even Poland decarbonizes faster than we do.

@jcoglan Yeah! Especially when the reviewed code is just snippets copied from random sites on the internet! Fun!

@janl@narrativ.es @mrchrisadams@mastodon.social Real-time collaboration and offline-first have one thing in common: potential order dependency of independent edits. This is what CRDTs solve.

@petealexharris @nicol @internetsdairy @TheBreadmonkey Could be more subtle, just number the last one “0.”

@johncarlosbaez@mathstodon.xyz Apparently not really free — I got Access Denied, probably because I block all the fingerprinting and tracking.

@CLEW@mastodon.energy I think it's a weak article. It almost seems to avoid explaining what went wrong.

Some facts that IMHO are cut short:

• Energiewende mainly replaced nuclear with gas
• Kohl presided over the stop of Germany's nuclear buildout, which until then was akin to France's Messmer plan
• Solar and wind at scale had an EROI < 1 before around 2000 (Energy Return On Investment)
• Greens promoted even coal over nuclear, effectively even in 2022.

@Lana@beige.party I wouldn't say that Obama's indecisiveness was the deciding factor for the gradual up-scaling of Russian aggression, but it definitely helped, and probably was a (for Russia) important part of the lackluster international reaction. I remember that many people said things like »yeah, OK, he's much better than the one before, but a /peace price/ seems a bit far fetched.« But of course, not nearly as far fetched as even this joke toy from fifa.

@FoxVK@mastodon.cloud @xChaos@f.cz The problem with promised profit is when it's a scam, and more so when hundreds of billions of dollars end up tied into that scam, and everyone is just waiting how bad the collapse will be. Cheating is when you pretend to achieve something that you don't. In Go, it's when you pretend to play the game, when you're just parroting a bot. In LLMs, it's when you pretend to create answers, when all you do is generate something that only looks like answers.

@infobeautiful@vis.social You should not use a link shortener. Here is the real link: https://informationisbeautiful.net/visualizations/common-mythconceptions/ Link shorteners are useless on Mastodon, and harmful for users.

@AstroKatie It fuels a hypothesis that I have been ruminating for a long time: what if the redshift is not a property of spacetime but of the light, i. e. it is losing (dissipating) energy ever so slightly over long distances?