Trevor Back

494 posts

Trevor Back

Trevor Back

@TrevorBack

MAI Multimodal - Product. Previously @GoogleDeepMind, @ShiftLabAI, @Speechmatics. Speech is core to the future AGI stack.

London Katılım Şubat 2011
1K Takip Edilen880 Takipçiler
Sabitlenmiş Tweet
Trevor Back
Trevor Back@TrevorBack·
Today, @MicrosoftAI is now 3rd on LMArena lab leaderboard with our latest model: MAI-Image-2. arena.ai/leaderboard/te… I’ve been playing with the model a lot these last few weeks and the capabilities are truly impressive.
English
1
2
8
237
YAROBONZ
YAROBONZ@URUBONZ_·
@mustafasuleyman @MicrosoftAI @arena Rate Limits and restrictions are to high, I generated 1 good image then struggled to do anything before getting rate limited for 24 hours
English
1
0
1
1.2K
Trevor Back retweetledi
Mustafa Suleyman
Mustafa Suleyman@mustafasuleyman·
Our new image generator MAI-Image-2 is out! Available now on MAI Playground for everything from lifelike realism to detailed infographics. Our team has been pushing immensely hard for this release, and we are now among the top models out there: #3 family on @arena. Check out the details in our blog: microsoft.ai/news/introduci… It's shipping soon in Copilot and Bing Image Creator, as well as Microsoft Foundry. Really proud of our progress on models and products - stay tuned for new releases and come join us on our Superintelligence mission!
Mustafa Suleyman tweet mediaMustafa Suleyman tweet mediaMustafa Suleyman tweet mediaMustafa Suleyman tweet media
English
77
92
497
168K
Microsoft AI
Microsoft AI@MicrosoftAI·
Meet MAI‑Image‑2. Built with creatives, for real creative work. Ranked #5 on @arena’s text‑to‑image leaderboard. Available now: msft.it/6014QUCBe
English
53
120
823
99.7K
Trevor Back
Trevor Back@TrevorBack·
Lovely graphic from @arena showing our progress since Image-1 🙏
Arena.ai@arena

Let’s dive deeper into the massive improvements between MAI-Image-2 vs. MAI-Image-1 by @MicrosoftAI. MAI-Image-2 shows significant gains across all sub-categories for Text-to-Image: Gains across all 7 sub-categories in order of magnitude: - Text Rendering (+115 pts) - Portraits (+105 pts) - Product, Branding & Commercial Design (+102 pts) - Photorealistic & Cinematic Imagery (+97 pts) - 3D Imaging & Modeling (+92 pts) - Art (+87 pts) - Cartoon, Anime & Fantasy (+81 pts)

English
0
0
3
114
Trevor Back retweetledi
Arena.ai
Arena.ai@arena·
Let’s dive deeper into the massive improvements between MAI-Image-2 vs. MAI-Image-1 by @MicrosoftAI. MAI-Image-2 shows significant gains across all sub-categories for Text-to-Image: Gains across all 7 sub-categories in order of magnitude: - Text Rendering (+115 pts) - Portraits (+105 pts) - Product, Branding & Commercial Design (+102 pts) - Photorealistic & Cinematic Imagery (+97 pts) - 3D Imaging & Modeling (+92 pts) - Art (+87 pts) - Cartoon, Anime & Fantasy (+81 pts)
Arena.ai tweet media
Arena.ai@arena

MAI-Image-2 debuts at #5 in the Image Arena! Highlights: - #5 in Text-to-Image overall - #5 for 3D Imaging & Modeling, Cartoon, Anime & Fantasy, Photorealistic & Cinematic Imagery, Art and Portraits - #6 for Product, Branding & Commercial Design Congrats to the @MicrosoftAI team on this milestone!

English
8
14
124
15.6K
Trevor Back
Trevor Back@TrevorBack·
And of course, there’s always space for a little bit of fun...
Trevor Back tweet media
English
1
0
0
32
Trevor Back
Trevor Back@TrevorBack·
Today, @MicrosoftAI is now 3rd on LMArena lab leaderboard with our latest model: MAI-Image-2. arena.ai/leaderboard/te… I’ve been playing with the model a lot these last few weeks and the capabilities are truly impressive.
English
1
2
8
237
Trevor Back
Trevor Back@TrevorBack·
Personalised AI Assistants don’t need to be as complex as you’d think. Sometimes great product sense and a deep understanding of the models delivers an amazing experience. Go check out this new product from my friend and ex-colleague @dougli_
Douglas Li@dougli_

🪴 Meet Lightfern - the telepathic AI writing tool. I left OpenAI to build Cursor for email. AI-driven communication isn’t authentic. You want to communicate clearly and thoughtfully. But wording your thoughts is time-consuming. Not anymore. We’ve developed the world’s best autocomplete model, proudly from Europe. Meet @lightfern_ai , the telepathic AI writing tool. It finishes your thoughts before your fingers hit the keys. Working directly in your inbox as a browser extension, Lightfern nails the details that make emails feel like you - nicknames, sign-offs, your usual tone - and pulls context from past threads to finish your sentences the way you would. Build stronger relationships. Lightfern keeps you thoughtful, whether it’s remembering your client’s holiday to New York or a long-forgotten invoice number. Revise and edit directly in your inbox. And no more copy pasting from emails to your LLM. Chat directly with Lightfern in your email inbox – it sees what you see. All this, plus zero data retention by default. Your emails, your data. It’s unmatched in contextual awareness and speed - one user even remarked that it’s “the closest thing to telepathy” she’s experienced. Human-first AI means human-first everything. That’s why our launch video is 100% hand-drawn: to put our money where our mouths are. 🪴Install for free today: lightfern.com. Lightfern is completely free during our beta period. P.S. If you’re in London, drop a comment / DM to get invited to our exclusive launch event!

English
1
0
4
240
Trevor Back
Trevor Back@TrevorBack·
We are still so early in speech AI... many assume users only want to speak with a single AI, and listen to a generated podcast of two speakers… but this isn’t how we interact with people IRL. In the near future I can’t wait to have richer dialogues with many AI personas
ElevenLabs@ElevenLabs

We are launching the Eleven v3 (alpha) API. Built for async use cases, Eleven v3 (alpha) delivers the most expressive Text to Speech model: - Dialogue mode, unlimited amount of speakers - 70+ languages - Enhanced voice and emotional control with [audio tags]

English
0
1
2
334