MAI-Thinking-1

(microsoft.ai)

100 points | by LER0ever 2 hours ago

11 comments

  • keeda 1 hour ago
    > Second, clean data. MAI-Thinking-1 was trained on clean and appropriately licensed data, with AI-generated content excluded from pre-training. This matters for quality, provenance, and control. If we cannot account for what shaped a model, we cannot fully understand its behavior or credibly improve it.

    Shots fired?

    It would be interesting to see how far "clean data" can go on the scaling laws.

    • onlyrealcuzzo 1 hour ago
      I'm interested how much "Clean Data" is synthetic data from "unclean" models...
      • bicx 17 minutes ago
        So, laundered data?
      • ertgbnm 52 minutes ago
        > with AI-generated content excluded from pre-training.

        > without distillation from third-party models

        sounds like zero unless they are lying.

        • postalcoder 13 minutes ago
          There’s a reason why they qualified it with a “pre-train”. Obviously they’re post training with synthetic data.

          Why are pre and post train ethically different, though?

        • zamalek 40 minutes ago
          > with AI-generated content excluded from pre-training.

          Though this is largely impossible these days, unless they pre-trained on pre-AI era data.

      • xavriley 1 hour ago
        “ We trained it from the ground up on enterprise grade, clean and commercially licensed data, without distillation from third-party models.”
        • azinman2 58 minutes ago
          aka all of GitHub OSS
    • vdfs 49 minutes ago
      I doubt any lab would say otherwise, they all _claim_ to use licensed data
      • keeda 30 minutes ago
        Maybe, but Microsoft, through their partnership with OpenAI, is already involved in major copyright lawsuits. That is probably a driving force for this move, actually... I doubt they would want to tempt fate while those lawsuits are on-going.
  • pixeldash928 2 hours ago
    Looks like the OAI divergence is finally taking place. Seems like the comparisons are mainly with Opus 4.6 and GPT 5.4 though. Still, exciting to see a new frontier player.
    • i_have_an_idea 35 minutes ago
      Is it a frontier player though, or perhaps a new benchmaxxed model? People were saying similar things about Grok but it ultimately amounted to little.
      • wasabi991011 16 minutes ago
        "preferred by humans over Sonnet 4.6" makes it pretty clearly not benchmaxxed though.

        At least when you define benchmaxxed as "good in benchmarks but not human preference".

  • hartator 15 minutes ago
    I like it so much when a website hijacks the way my scroll works. This is truly innovative.
  • BeetleB 32 minutes ago
    Based on the first table, why would I pick this over GLM?
    • missedthecue 11 minutes ago
      Because your employer might make you exclusively use enterprise copilot.
  • lordmauve 1 hour ago
    We need to see DeepSWE scores. SWE Bench Pro is junk.
  • kstenerud 51 minutes ago
    They've hijacked scrolling. They've hijacked the spacebar. It flickers like crazy when I try to move through the article. Trying to get through it is an exercise in madness.
    • t-sauer 35 minutes ago
      I do not understand how scroll hijacking is still a thing. Who thinks this is a better experience?
    • grassfedgeek 9 minutes ago
      Even without flicker it is very distracting. Why do people think this is a good idea?
    • AirMax98 47 minutes ago
      I normally don't comment on matters of taste like this, but wow this is brutal. It's like someone threw the site in a vat of molasses.
    • blisstonia 19 minutes ago
      I gave up after the first scroll.
    • aniceperson 34 minutes ago
      there is also a gap between the header and the top of the page... they should ask the ai to make it better a few more times...
  • vcryan 12 minutes ago
    It really looks like they used Claude to design this webpage. I guess the color taupe it the marker of good AI today.
  • wmf 1 hour ago
    At least there shouldn't be any complaints about benchmaxing this time.
    • i_have_an_idea 34 minutes ago
      Just because it is performing rather poorly by comparison, it doesn’t mean it isn’t benchmaxxed. It can still be worse than it appears.
      • wasabi991011 15 minutes ago
        It isn't benchmaxxed because they are using human preference as an evaluation.
  • bossyTeacher 1 hour ago
    7 modes launched. 5 models in the dropdown. Only 4 actually usable :(

    About time Microsoft joined the fray. After the OpenAI divorce, it really looked like Microsoft was going to become another Uber.

    • giancarlostoro 1 hour ago
      They still own 27% of OpenAI, this IPO will feed them a lot of easy cash.
  • simjnd 2 hours ago
    Absolutely disgusting scroll jacking, even when "Accessibility mode" is turned on
    • dang 1 hour ago
      I'm sure most of us agree, but:

      "Please don't complain about tangential annoyances—e.g. article or website formats, name collisions, or back-button breakage. They're too common to be interesting."

      https://news.ycombinator.com/newsguidelines.html

      • simjnd 32 minutes ago
        Forgot about this, my bad!