MAI-Thinking-1 • Hacker News Board

Re: MAI-Thinking-1

#81

Post by inquirerGeneral » Wed, Jun 03, 2026, 3:20 AM UTC

supermdguy wrote: post #55

keeda wrote: post #5

> 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.

It's interesting because their last model series (Phi) was based around the thesis that high-quality synthetic data is better than a large pre-training corpus.

[dead]

Re: MAI-Thinking-1

#82

Post by rurban » Wed, Jun 03, 2026, 4:47 AM UTC

azinman2 wrote: post #12

xavriley wrote: post #7

Earlier quoted context omitted.

“ We trained it from the ground up on enterprise grade, clean and commercially licensed data, without distillation from third-party models.”

aka all of GitHub OSS

Not OSS only, likely also the enterprise private repos, with a lot of business secrets.

Re: MAI-Thinking-1

#83

Post by aesthesia » Wed, Jun 03, 2026, 5:44 AM UTC

What's interesting is that although they don't seem to be releasing the model weights, they have published a technical report (https://microsoft.ai/wp-content/uploads/2026/06/main_2026060...) that's more extensive than the typical open-weights model gets.

Re: MAI-Thinking-1

#84

Post by dr_kiszonka » Wed, Jun 03, 2026, 8:41 AM UTC

Centigonal wrote: post #41

> MAI-Thinking-1 is a 35B-active, ~1T-total parameters, sparse Mixture of Experts model, a smaller inference footprint than much larger models. This seemingly nonsensical sentence (of course this will have a smaller inference footprint than larger models) suggests this model's competitors have larger inference footprints and total parameter sizes.

When would a larger model have a smaller inference footprint? If the larger was MoE and the smaller was dense?

Re: MAI-Thinking-1

#85

Post by Leynos » Wed, Jun 03, 2026, 8:45 AM UTC

chongli wrote: post #50

saghm wrote: post #42

Earlier quoted context omitted.

"how many of those shapes are rectangles?" "sounds like zero unless they are squares" Adding "unless" to a statement makes it vacuous if the latter clause is weaker than the first clause. I find it hard to believe that a company willing to violate licenses would have scruples about lying about it.

Adding "unless" to a statement makes it vacuous if the latter clause is weaker than the first clause I think that's the point. "How do I say they're lying without outright saying they're lying?" It's a common rhetorical trick.

Or the speaker is just not in the mood to argue with someone whose response will be, "you trust anything Microsoft say?"

Re: MAI-Thinking-1

#86

Post by Bolwin » Wed, Jun 03, 2026, 3:34 PM UTC

vb-8448 wrote: post #54

Alifatisk wrote: post #38

> MAI-Thinking-1 is built with enterprise readiness in mind. It supports long context with a 256k token window Isn’t 1M becoming the norm?

1M it's only marketing, in my experience above 150k quality noticeable drops. Claude code will suggest you to start a new session or compact if you go above 100k.

In my experience above 60k quality noticeably drops.

30k for open source models

Re: MAI-Thinking-1

#87

Post by Centigonal » Wed, Jun 03, 2026, 4:10 PM UTC

dr_kiszonka wrote: post #84

Centigonal wrote: post #41

> MAI-Thinking-1 is a 35B-active, ~1T-total parameters, sparse Mixture of Experts model, a smaller inference footprint than much larger models. This seemingly nonsensical sentence (of course this will have a smaller inference footprint than larger models) suggests this model's competitors have larger inference footprints and total parameter sizes.

When would a larger model have a smaller inference footprint? If the larger was MoE and the smaller was dense?

yes, MoE reduces the inference compute requirements (inference memory reqs remain the same)

Re: MAI-Thinking-1

#88

Post by stingraycharles » Thu, Jun 04, 2026, 12:55 AM UTC

foresterre wrote: post #68

stingraycharles wrote: post #45

Earlier quoted context omitted.

I assume they took the actual repos’ licenses info account. I don’t understand why they should ask for permission when the license would already allow for it.

Almost all licenses have requirements to redistribute copies of the work, or derivatives thereof. Even permissive licenses do. It's very little to ask when open source dev's provided thousands of hours of free work. For example, the Apache 2.0 license requires in just 4.c: You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from th…

But that’s what I meant with taking it into account. They would likely only use BSD and MIT licensed repos, which is a lot.

Re: MAI-Thinking-1

#89

Post by rajveerb » Thu, Jun 04, 2026, 3:13 AM UTC

Centigonal wrote: post #87

dr_kiszonka wrote: post #84

Earlier quoted context omitted.

When would a larger model have a smaller inference footprint? If the larger was MoE and the smaller was dense?

yes, MoE reduces the inference compute requirements (inference memory reqs remain the same)

As someone who has spent quite a lot of time on inference, I would a add a small note:

Deployment looks very different for MoE than dense style models so I would say that it is more nuanced than "inference memory reqs remain the same". Memory can be very different for MoE style models.

Re: MAI-Thinking-1

#90

Post by deflator » Thu, Jun 04, 2026, 1:38 PM UTC

Does this mean that work created with it can be copyrighted? Since the courts ruled that the inclusion of pilfered IP was the reason other model's work cannot be copyrighted, I would think so! In that case, this is a completely different beast. It can maybe be used for things that need a durable copyright.