armchairhacker
Or they could explicitly promote Lula, but that would be biased towards him…
LLMs make what was formerly hard trivial, that is true. But they cannot do everything with simple prompts, you must increase your ambition to what was formerly impossible. Try building a bigger game, more complicated software or hardware…what have you always wanted but written off as something only a multi-millionaire with a large team of domain experts could accomplish? Because you have one now.
“End algorithms”? Even a firehouse of newest comment (or most upvoted as suggested) would benefit Bolsonaro if his online support outnumbers opposition. Ban both candidates from the platform, or make every search show them both? That will probably not work; confuse Brazilian citizens why their politicians are being censored (which will influence them because their politician is being attacked by Lula’s court, or even Facebook acknowledges their opponent politician is dangerous); and not end the lawsuits if they continue bringing “evidence” like Meta AI randomly generating Bolsonaro-related numbers.
The most egregious evidence against Meta are suspending Lula’s official accounts. This can also be attributed to a neutral system processing opponents’ mass reports, but has an obvious solution: give every official account immunity from suspension except by the highest level admins, because they’re supposed to only be suspected according to guidelines, and a $trillion company has enough resources to closely monitor high-level official accounts.
In your other comment, “Talking also costs.”, it’s clarified through context that talking costs tokens.
I think it’s a side-effect of the AI companies trying to reduce verbosity and redundancy, which were my personal main complaints with LLM writing. Though I believe LLMs are still verbose, and I suspect it’s because they don’t know what to remove, and these are cheap gains where the LLM removes verbosity that is obvious but expected because everyone does it.
The LLMs did their job (except giving the same answer demonstrated their toxic homogeneity).
Remember when AI was extremely sycophantic (as opposed to now when, one might say, it’s mildly sycophantic)?
Last I’ve heard, most civilians polled want the war to end…but would rather keep fighting than end without concessions Russia isn’t willing (IIRC restoring occupied territory, maybe NATO). But I think that was over a year ago.
Personally (I a non-Ukrainian) think it’s horrendous they physically kidnap young men and place them in the front lines, which they are accused of and probably to some extent doing. I doubt it’s effective, and if it’s that or lose, they’ve already lost. But if the Ukrainian majority in Ukraine does not want to officially surrender, and I can definitely understand why, we should keep supporting them. Otherwise it should end. Though if denying weapons and other resources is unlikely to reduce the bloodshed, we should continue providing them but change our messaging, using that evidence to persuade other nations.
Society is very large, so we can tackle big and small issues. And I’m sure it’s more effective to let people choose smaller issues, than force them to work on big issues they don’t care about and are probably redundant.
Models are good at recreating existing projects that already have existing companies.
Ok, they are good at porting and combining existing things. But anyone can run OpenAI's or Anthropic's models themselves, or if both cooperate and significantly raise prices, I bet businesses and consumers would just wait until their competitors catch up, and a significant amount would permanently switch.
The first step, arresting the most obvious, won’t lead to a Reign of Terror on its own. But I don’t think it will be effective either.
Seriously, Mark Zuckerburg doesn't really care that one person left his platform, but he actually would prefer you stay...because if enough do it will affect earnings. But you should leave Facebook not to upset Zuck, but to stop upsetting yourself with what's on there.
You can’t ban ideas, because they will hide behind other ideas, until you ban everything. If the AfD is banned and voters start holding blank pieces of paper and posting *** like in Russia, everyone knows what they mean, but will you start arresting people holding blank pieces of paper and posting *** like in Russia?
Ultimately, you can’t ban ideas in a democracy, unfortunately even the idea to abolish it. What you can do it make that process harder and slower, so your opponents lose power before they succeed. Or, you know, convince people to drop the idea altogether.
I actually disagree with the former, and wholly agree that AI should be allowed in Linux. We want to measure AI’s growing contribution, demonstrate that it can write good code, and identify problems that AI code that don’t appear in human code so they can be fixed.
> Jeff Layton, a kernel developer at Red Hat, had earlier given three reasons to drop it. Not everyone follows the policy, he said. The tag “basically serves as free advertising for proprietary LLM companies”, he added, and it is unclear why the kernel collects the information at all.
Even though not everyone follows the policy it’s still roughly accurate and maintainers can remove the model company from the commits.
The pessimist says creative projects are rare nowadays, an optimist says they're still coming out. You may also dislike the vibe-coded "...Craft" projects, but seemingly building a production-capable software in days is something that has never been done before and IMO inspiring.