[–] 1 point 2 months ago

I don't think Randall Munroe would argue others are dumber than he is, only that he's seeing a scenario that is not (as frequently) reflected in fiction and thought experiments as the classic AI gains sentience, escapes containment and takes over scenario.

I'd point to Linus' Law, Given enough eyes, all bugs are shallow. (Wikipedia notes the law's validity is confirmed empirically, but refuted in analysis.) This applies to more than software. Right now, election enthusiasts in the US are combing election law and procedure for ways the November election can be thrown, and ways to counter that scenario quickly.

I was thinking about the misprogramming scenario of AI escape. This implies hyperintelligent AI is dangerous even when everyone using it is benign and has non-adversarial intentions. Essentially an AI will need to be coded with an ethics system such as that of a general (or high-ranking soldier in the US) who deliberates on operational action and its consequences, whether the operation is legal, or would cause undue harm in its execution.

Right now, AI in its current development is... whimsical, at best. That is to say, it ignores rules the way it hallucinates facts. And AI can be manipulated to ignore rules through adversarial input, so AI is not ready to operate dangerous machinery.

  • source
  • parent
  • context
  • [–] 4 points 2 months ago (1 child)

    'Finishing' a robot in 10 mins.

    MAKING a robot in 10 mins is insane. But I guess if they are adding those robots to the factory lines, something something 'many hands make light work'.

  • source
  • [–] 3 points 2 months ago

    Could be either honestly. I always hated how AWS did their multi-region setup, where it's basically a whole separate account. So difficult to do multi region setups. GCP and Azure both made that much much easier. For this exact reason, I don't really trust AZs. They're all too close together still. I want the data replicated around the planet.

  • source
  • parent
  • context
  • [–] 5 points 2 months ago

    There is no rational humanitarian choice between the two futures. The 2077 scenario has to go away. Simple as that.

    I, we, have to reject, boycott and fight all big proprietary AI Corporations and support, so they can't create the 2077 - or 'Blade Runner' - version of our future. Global Communism is the natural state of living together when we can design the future we want. We can produce anything we need in an intelligent circular eco-system instead of hoarding/fighting for artificially scarce resources, while thrashing our whole planetary habitat at the same time. We can setup the environment and systems for everyone to prosper and flourish with the passions they have.

    We no longer need to 'compete' at all cost, and we are not in a hurry for anything except solving the problems our current system generate. We don't need to fight against everyone else for our basic livelihood and security - we just need common access to shared resources. We can do better than insects and other low-cognition organisms that can't fathom, or redesign their world, and thus need to compete for survival - all the time..

    Let's choose not to behave like that any more.

    I definitely support Deepseek in its cause, and reject proprietary Corporations and their narrow-minded insect-like doctrines and visions for our common future..

  • source
  • [–] 22 points 2 months ago* (last edited 2 months ago) (1 child)

    The whole point of AWS Availability Zones (AZ) is that if one location goes out, you still have data elsewhere and it would automagically recover. The default for services in one Region is 3 AZs. Some regions like us-east-1 have 6 AZs.

    https://aws.amazon.com/about-aws/global-infrastructure/regions_az/

    From their own map, it looks like both Bahrain and UAE had 3 AZs.

    https://builder.aws.com/content/3ADBRZtN2DSQEgLOa3ga88M4dnx/how-many-azs-you-need-to-deploy-to-in-prod-environment

    Amazon also recommends storing data across at least 3 AZs to avoid data loss.

    So either customers actively turned off multi-AZ support, someone screwed up the local installation, or EVERY single one of those AZs got hit.

    Edit: article is paywalled, but there's a small section that says multiple AWS locations were hit in Bahrain.

  • source
  • [–] 4 points 2 months ago

    That's not good as good news as the title makes it sound.

    According to the analysis, anyone walking through the city with the glasses would have to inform passers-by verbally in advance or even wear high-visibility vests with inscriptions to comply with the rulings of the European Court of Justice. Particularly sensitive areas include doctor’s offices, demonstrations, women’s shelters, or playgrounds. The use of Ray-Ban Meta AI Glasses in public spaces is therefore difficult to make legally compliant from a data protection perspective.

    Since it's up to the individuals to disclose, they're the ones at fault. So you'd have to know you were being recorded, and by who, and then hold them accountable. That's pretty much impossible.

    But more importantly, meta is no longer held responsible, and they can just continue fucking up people's privacy.

  • source
  • [–] 10 points 2 months ago

    "I certainly hope I won’t be revolutionized, but if I must be revolutionized, I hope the one revolutionizing me is myself.” When everyone is so obsessed with destroying themselves, I have to join this cruel arms race as well.

    Some real fight club vibes there.

  • source
  • view more: ‹ prev next ›