Dave5

About

Tag: AI/ML

  • No, LLMs still can’t reason like humans. This simple test reveals why.

    So, what explains the shocking failures of basic reasoning that Simple Bench exposes? Let’s return to the vegetables — or fruits and vegetables, if you wish. Here’s my take on the first of my two questions: Why do models fail the question you saw above? … The clue is in their name: language models. They model language.…

  • “Self-Driving” Cars Have a Dirty Secret

    It’d be remiss to fault robotaxi companies for having these kinds of safety fallbacks. The problem is that they’re not upfront about the fact that humans are still in the loop, creating a facade of full autonomy that doesn’t actually exist. https://futurism.com/the-byte/self-driving-dirty-secret

  • Snake Oil, Baby

    src: https://www.linkedin.com/posts/denis-o-b61a379a_genai-activity-7237647652052488192-CwHF?utm_source=share&utm_medium=member_desktop

  • The Tesla Workers Dealing With Autopilot Data Are Treated So Strangely You Won’t Believe It

    With new headlines cropping up all the time showing how poorly a job Tesla Autopilot actually does at piloting cars, it’s not surprising that the people tasked with building that system’s knowledge out are underpaid and heavily surveilled — or that when they tried to fight back against what they felt was unfair treatment, they were summarily cast aside. https://futurism.com/the-byte/tesla-autopilot-workers-treatment

  • Not the Onion: Telsa FSD doesn’t even work in a tunnel

    Even more… it doesn’t even work in a tunnel that Telsa has full environmental control of. Forget open roads. Tesla still hasn’t figured out how to get its vehicles to autonomously drive through Elon Musk’s tunneling venture The Boring Company’s single-lane tunnel. https://futurism.com/the-byte/tesla-full-self-driving-doesnt-work-tunnel So explain to me why this fucker is allowed to put this…

  • Copilot AI calls journalist a child abuser, Microsoft tries to launder responsibility

    Martin Bernklau is a German journalist who reported for decades on criminal trials. He looked himself up on Bing, which suggests you use its Copilot AI. Copilot then listed a string of crimes Bernklau had supposedly committed — saying that he was an abusive undertaker exploiting widows, a child abuser, an escaped criminal mental patient.

  • AI Appears to Be Slowly Killing Itself

    “The web is becoming increasingly a dangerous place to look for your data,” Rice University graduate student Sina Alemohammad, who coauthored a 2023 paper that coined the term “MAD” — short for “Model Autophagy Disorder” — to describe the effects of AI self-consumption, told the NYT. https://futurism.com/ai-slowly-killing-itself

  • Māori and First Nations Australia musicians views of Artificial Intelligence (AI)

    From: https://www.linkedin.com/pulse/m%25C4%2581ori-first-nations-australia-musicians-views-ai-taiuru-jp-minstd-klacc/?trackingId=L6r7lk1ZfhkZb6MVmf9fJg%3D%3D Source report: https://www.apraamcos.co.nz/about-us/news-and-events/ai-and-music-survey-results

  • Why AI’s Tom Cruise problem means it is ‘doomed to fail’

    One way to explain this is to realise that LLMs don’t learn about relationships between facts, but between tokens, the linguistic forms that Bender described. The tokens “Tom Cruise’s mother” are linked to the tokens “Mary Lee Pfeiffer”, but the reverse is not necessarily true. The model isn’t reasoning, it’s playing with words, and the fact…

  • AI Unicorns Are Running Amok

    As someone who has been around this world of technology, startups and venture capital (in many different roles), let me tell you that there isn’t a new trend venture capitalists don’t love. New trends mean new companies to back and hope that there is a winner or two in those bets. It’s a good way…