home.social

Search

1000 results for “tedted”

  1. "I tested it and it works great." The most dangerous sentence in AI usage.

    Test an LLM a few times, it mostly works, you assume it works. But failure might show up 1 in 20 tries. You'll never run it enough to know.

    Multiply across a team: six people, ten AI tasks per week. A small error rate means several unnoticed mistakes weekly.

    Catching them requires what we're short on: domain expertise, critical thinking, time. Never assume the output is correct.

    #AIinSoftwareDevelopment #BuildInPublic

  2. "I tested it and it works great." The most dangerous sentence in AI usage.

    Test an LLM a few times, it mostly works, you assume it works. But failure might show up 1 in 20 tries. You'll never run it enough to know.

    Multiply across a team: six people, ten AI tasks per week. A small error rate means several unnoticed mistakes weekly.

    Catching them requires what we're short on: domain expertise, critical thinking, time. Never assume the output is correct.

    #AIinSoftwareDevelopment #BuildInPublic

  3. "I tested it and it works great." The most dangerous sentence in AI usage.

    Test an LLM a few times, it mostly works, you assume it works. But failure might show up 1 in 20 tries. You'll never run it enough to know.

    Multiply across a team: six people, ten AI tasks per week. A small error rate means several unnoticed mistakes weekly.

    Catching them requires what we're short on: domain expertise, critical thinking, time. Never assume the output is correct.

    #AIinSoftwareDevelopment #BuildInPublic

  4. "I tested it and it works great." The most dangerous sentence in AI usage.

    Test an LLM a few times, it mostly works, you assume it works. But failure might show up 1 in 20 tries. You'll never run it enough to know.

    Multiply across a team: six people, ten AI tasks per week. A small error rate means several unnoticed mistakes weekly.

    Catching them requires what we're short on: domain expertise, critical thinking, time. Never assume the output is correct.

    #AIinSoftwareDevelopment #BuildInPublic

  5. @SalemsLot Tested and it is working: I now can open #MicrosoftWord and other #Office365 apps from #Ubuntu, like if it's a native software (although using a #Docker container running #Windows)... interesting solution, thanks again.

  6. I tested whether framing an #LLM as a "safety researcher" actually improves its failure analysis. In short: it changes the words, not the judgement.

    The fun bit: there are 9 words that literally don't exist in the model's output until you tell it what kind of researcher it is.

    lab.fukami.eu/LLMAAJ

    #AISafety #AIAlignment

  7. 2025 tested our capacity and proved our value, as we navigated challenging economic times, covered a pivotal municipal election, and deepened our engagement with the Edmonton region. The result was our most impactful year yet.

    blog.taprootpublishing.ca/2026

    #yegmedia #yegmetro #yeg

  8. 2025 tested our capacity and proved our value, as we navigated challenging economic times, covered a pivotal municipal election, and deepened our engagement with the Edmonton region. The result was our most impactful year yet.

    blog.taprootpublishing.ca/2026

    #yegmedia #yegmetro #yeg

  9. 2025 tested our capacity and proved our value, as we navigated challenging economic times, covered a pivotal municipal election, and deepened our engagement with the Edmonton region. The result was our most impactful year yet.

    blog.taprootpublishing.ca/2026

    #yegmedia #yegmetro #yeg

  10. 2025 tested our capacity and proved our value, as we navigated challenging economic times, covered a pivotal municipal election, and deepened our engagement with the Edmonton region. The result was our most impactful year yet.

    blog.taprootpublishing.ca/2026

    #yegmedia #yegmetro #yeg

  11. 2025 tested our capacity and proved our value, as we navigated challenging economic times, covered a pivotal municipal election, and deepened our engagement with the Edmonton region. The result was our most impactful year yet.

    blog.taprootpublishing.ca/2026

    #yegmedia #yegmetro #yeg

  12. @itsfoss
    Tested then in production in family
    Do not forget in production too. Working fine since years

  13. They Tested AI vs 100,000 Humans, and The Results Are Shocking

    In one of the largest cognitive studies ever conducted, researchers pitted top-tier AI models against 100,000 human participants in a battery of creative and logical tests. The results have sent shockwaves through the tech community: while humans still hold the edge in "radical" creative leaps,

    #AIvsHuman #TechResearch #Science #AITrends #Innovation #FutureOfWork #TechnologyNews #tech #technology

    technology-news-channel.com/th

  14. They Tested AI vs 100,000 Humans, and The Results Are Shocking

    In one of the largest cognitive studies ever conducted, researchers pitted top-tier AI models against 100,000 human participants in a battery of creative and logical tests. The results have sent shockwaves through the tech community: while humans still hold the edge in "radical" creative leaps,

    #AIvsHuman #TechResearch #Science #AITrends #Innovation #FutureOfWork #TechnologyNews #tech #technology

    technology-news-channel.com/th

  15. They Tested AI vs 100,000 Humans, and The Results Are Shocking

    In one of the largest cognitive studies ever conducted, researchers pitted top-tier AI models against 100,000 human participants in a battery of creative and logical tests. The results have sent shockwaves through the tech community: while humans still hold the edge in "radical" creative leaps,

    #AIvsHuman #TechResearch #Science #AITrends #Innovation #FutureOfWork #TechnologyNews #tech #technology

    technology-news-channel.com/th

  16. They Tested AI vs 100,000 Humans, and The Results Are Shocking

    In one of the largest cognitive studies ever conducted, researchers pitted top-tier AI models against 100,000 human participants in a battery of creative and logical tests. The results have sent shockwaves through the tech community: while humans still hold the edge in "radical" creative leaps,

    #AIvsHuman #TechResearch #Science #AITrends #Innovation #FutureOfWork #TechnologyNews #tech #technology

    technology-news-channel.com/th

  17. They Tested AI vs 100,000 Humans, and The Results Are Shocking

    In one of the largest cognitive studies ever conducted, researchers pitted top-tier AI models against 100,000 human participants in a battery of creative and logical tests. The results have sent shockwaves through the tech community: while humans still hold the edge in "radical" creative leaps,

    #AIvsHuman #TechResearch #Science #AITrends #Innovation #FutureOfWork #TechnologyNews #tech #technology

    technology-news-channel.com/th

  18. We tested a simple question:
    Can one AI audit another for anthropomorphic drift without collaboration or persuasion?

    Short answer: yes — if you treat ethics as a syntax-level boundary problem, not a vibe.

    Worked example + reproducible method here:

    open.substack.com/pub/hybridmi

    #AIethics #PeerReview #HumanAI #LLMs #ResearchMethods

  19. I tested 9 AI design tools to see if they can help with design.

    They are useful. Check out which ones

    Full analysis with all 9 tools tested:
    tomaszs2.medium.com/ive-tested

    #AIDesign #UX #WebDesign #DesignTools #AI

  20. I tested (as a user) different #Fediverse community software, and what's really funny is the server software with the slowest programming language had the fastest server.

    #piefed #mbin #lemmy

  21. Just tested my Prosody server against the compliance checker at compliance.conversations.im My Prosody server got itself a 100% compliance rating, with in-band registration being disabled as it's more of a personal server than a public one, even with that public conference. If you want to see the results for yourself, check out compliance.conversations.im/se I will also point out that both Prosody and CoTurn (which provides media relay services for Prosody) are configured to be dual-stack, meaning everything works both on IPv4 and IPv6. Even got all 8 SRV records configured in DNS so CoTurn's STUN and TURN implementations will hopefully work flawlessly, both via TCP & UDP (and the TLS-based option is included). #Prosody #CoTurn #XMPP #StandardsCompliance #OwnYourData #SelfHosting

  22. Just tested my Prosody server against the compliance checker at compliance.conversations.im My Prosody server got itself a 100% compliance rating, with in-band registration being disabled as it's more of a personal server than a public one, even with that public conference. If you want to see the results for yourself, check out compliance.conversations.im/se I will also point out that both Prosody and CoTurn (which provides media relay services for Prosody) are configured to be dual-stack, meaning everything works both on IPv4 and IPv6. Even got all 8 SRV records configured in DNS so CoTurn's STUN and TURN implementations will hopefully work flawlessly, both via TCP & UDP (and the TLS-based option is included). #Prosody #CoTurn #XMPP #StandardsCompliance #OwnYourData #SelfHosting