Close Menu
    Politik

    Meinung: OpenAI hat gezeigt, dass man ihm nicht vertrauen kann. Kanada braucht eine verstaatlichte, öffentliche KI


    Meinung: OpenAI hat gezeigt, dass man ihm nicht vertrauen kann. Kanada braucht eine verstaatlichte, öffentliche KI

    https://www.theglobeandmail.com/business/commentary/article-openai-tumbler-ridge-chatgpt/?utm_source=dlvr.it&utm_medium=twitter

    View 14 Comments

    14 Kommentare

    1. varitok on

      We don’t need any. You’d have to be stupid to want a government to have that much information gathering potential on it’s own citizens.

    2. EuropesWeirdestKing on

      We need sovereign AI. Nationalized AI will not keep up. We can’t even run a post office, there is not a snowballs chance in hell we can run AI lol

      By the way there are sovereign AI solutions being developed in Canada

    3. BertramPotts on

      Big yes on public AI. AI is actually a useful technology, in much the same way spell check was a useful innovation. Custom built AI systems in particular have a lot of potential, while the AGI quest is just magical thinking led by some of the biggest grifters in history.

      The vision of AI where it needs to produce a consumer product that will generate 100s of billions of dollars of revenue for a few American corporations is the problem, it is also what all the hype train about AI is in service to.

      Public AI is a very different, very practical way of building AI into our lives in ways that actually work, it is how any sensible society would go about incorporating the technology, but because we live in the nadir of neoliberal civilization the only way we can think to promote AI is a top-down effort to shove a one-size-fits-all frankesntein as the solution to all problems.

    4. reward72 on

      LOL. How can someone follow what is happening in the US and say that the government should nationalize AI in the same sentence?! Aren’t we learning anything?!

      Big corporations cannot be trusted, but neither can governments. Greedy billionaires can lobby laws in their favour, but mad politicians can make the laws. It’s even worse.

    5. lavalamp360 on

      I’ll take a nationalized energy program and internet service provider before I want anything to do with AI infrastructure.

    6. DaOffensiveChicken on

      Im amazed people can see the government utterly fail in developing software like arrivecan or the phoenix pay systen and then still say developing this in house is a good idea lol

    7. kingbain on

      Since anthropic has fallen out of high fashion with the Whitehouse and is considered one if the best. Why not invite them over to the great white north and free.

    8. garden_gnorm on

      Federal Government should not run it, but they should invest heavily in Higher Ed and Research institutions to both work on developing AI, as well as providing the necessary funding for other areas of research to ride the coattails effectively.

      Canadian companies developing AI products in the healthcare, infrastructure, agriculture, education, etc sectors should receive grants and/or tax credits.

      If you incentive developing real solutions the hope is you end up with real results.

    9. Standard_Program7042 on

      They can’t be serious? the government struggled with a system to pay employees, a basic arrival app during covid, or even the CBC ability to handle streaming demand during the Olympics.

    10. Happyman321 on

      As with every other nationalized program it will be slow, outdated, filled with bloat, and cost us a fortune.

      No thanks

    11. jonlmbs on

      Spending millions (potentially billions) of dollars to create an inferior product that is instantly behind the state of the art seems like a terrible idea.

      Nationalize the data centres and infrastructure where models are running instead.

      Models will eventually be a commodity and open source models will be as good as commercial models. Controlling the infrastructure is more important.

    12. mummified_cosmonaut on

      In terms of culpability for the Tumbler Ridge massacre… Open AI is not very high on the list relative to the failures of law enforcement and the mental health system.

      The killer had done enough to have been locked away in a psych ward years ago.

    13. iamnotparanoid on

      We cannot trust AI companies, that is entirely true.

      What exactly would a nationalized AI company do for us, though? How would we benefit? The current AI business model is to allow companies to shrink their staff, and it’s doing it poorly.

      AI advocates are saying it’s the next big thing, but nobody has found a practical use for it yet that justifies the billions upon billions it requires in investments.

    14. Trid1977 on

      Shortly after graduating in Computer Science in the early 1980s there was a plan to certify Programmers like Engineers. Once networking became more prevalent there were protocols suggested that IT professionals would be obliged to report porn searches by users.

      Both these seemed to not get ‘implemented’. Although I did report porn searches to HR. If it’s suggested programmers report illegal activity, AI should have the same thing.

    Leave A Reply