Elektrine lite

← Feed

Peter Bloem

pbloem@mathstodon.xyz

<p>Assistant prof. at the Learning and Reasoning group, Vrije Universiteit Amsterdam.</p><p>Not an AI pusher, I promise. I just think we need a better class of criticism.</p>

Posts

  • View post

    Holy forking shirtballs. https://kingy.ai/blog/openai-ai-agents-cpu-cache-communication/

  • View post

    How long until AI unintentionally*, directly causes a single incident in which (a) at least hundreds of people die or (b) at least tens of thousands of people suffer something of comparable magnitude to losing your life&amp;#39;s savings? * Not intended by any human being. It may be that the AI itself has some vague notion of intention or that it just craps out in some random way.

  • View post

    RE: https://journa.host/@timkmak/117519148497991690 Elon Musk really is spectacular at incentivising people to put him out of business. Sometimes they slack off and start depending on his products, but he will say &amp;quot;no&amp;quot; and throttle their access, or break something crucial. He won&amp;#39;t let them sleep on the job. Build those alternatives and make them cheaper and more reliable. We need those competing product yesterday, people. He&amp;#39;s a harsh taskmaster, but he re...

  • View post

    &amp;quot;I’m not a betting man, but I have a prediction. It’s going to turn out that P=NP in some galactic sense, and all we’ll learn is that we spent fifty years answering the wrong question.&amp;quot; https://www.argmin.net/p/galactus-the-devourer

  • View post

    Autonomous drones following running people through the woods, without GPS. Developed, with military funds, at a public university. This is not dual use anymore, this is single use. There is only one use here. Sometimes I worry that I&#39;m becoming a negative Nancy with all my ethics talk, and then I see colleagues proudly post this kind of thing without a single acknowledgment of the implications.