Lynn Cherny

@venturetwins We want to see the pets

Thanks all for the incredibly helpful tips! And please keep them coming - I’ll upgrade for future videos. So far, I’ve realized everything I thought I knew about lighting was wrong 😅 First video will be going live tomorrow here: @justinemooreai" target="_blank" rel="nofollow noopener">youtube.com/@justinemooreai

Anthropic onboarding day: Michael Scott introducing Karpathy like he just signed Wemby in free agency.

@multimodalart It’s pretty good.

Back! huggingface.co/microsoft/Lens

Microsoft Lens is back! The 3.8B text-to-image model with GPT-OSS as the text-encoder. It was pulled from @huggingface last week, and it's back on all its might Playing with it here, join me on the vibe check huggingface.co/spaces/multimo…

I'm looking to find people interested in semi-supervised clustering, particularly with hard "must not be in the same cluster" constraints, to better understand your use cases.

@MartinNebelong Did it lose the gun in the last clip? Is that a safety filter problem?

I had to do this of course - test out the new Google Omni Flash model with my favorite 3d art sandbox - Dreams. Just a quick test for now.. I'll be sharing more examples and techniques soon but wanted to show you this one. There's something very fitting about combining "handmade" 3d models and animations sculpted from scratch in software that runs on 2013 hardware, with some of the most advanced AI technology in the world. As part of my work for Google, I've used Google V2V models on a few different projects now - Ancestra, Dear Upstairs Neighbors and most recently one of the IO films (I'll share more about that project soon) 🥰 - and video continues to be my favorite form of input for Generative Ai workflows. If an image says more than a thousand words, a video says more than a million. And knowing that these capabilities are now in the hands of all of you, makes me very excited to see the new workflows and projects that will come out of it. What do you think of Omni so far?

Stochastic Labs is running their summer residency again! This year's theme is Liveness + Agency — subgroups on artificial life, realtime AI interfaces, theater + code, experimental narrative interfaces, and more. Wonderful mix of artists / scientists / technical dreamers. Highly recommend. Apply ↓ stochasticlabs.org/liveness-agency

Today is the 10 year anniversary of Alt-AI! Alt-AI was a 3-day conference in NYC hosted by @sfpc and curated by me, exploring the then very niche field of creative AI. Lots of amazing artists and people participated, including talks by @mtyka @hdeweyh @RebeccaFiebrink @quasimondo @aparrish @golan @arnicas @bwhitman. There was also an exhibition with early AI art by @kcimc @memoakten @dribnet @hardmaru @schwittlick_ @xululululuuum and many others. I had never thought about organizing or curating before, but when @zachlieberman asked me to put it together, I thought I might never get another chance to put a roster like this into one room together, so I went for it. I'm still grateful and nostalgic for these days when this technology was so new weird and unexplored.

Join us as a Forward Deployed Engineer!

@wildmindai Huh is there code?

CVG: Compositional video gen via inference-time guidance. - no retraining - multi-stage actions - boosts Wan2.2/CogVideoX - improves motion direction, spatial relations, numeracy - Qwen2.5-VL for automated grounding It only steers the first 8 steps of the process, you get the structure you want, no sacrificing speed/img quality. arielshaulov.github.io/CVG/

Mid-month newsletter, with testing image-blaster for Claude/@theworldlabs, three.js games, weird science, LLMs behaving weirdly, and the usual other fun. A worker faery story on top. open.substack.com/pub/arnicas/p/…

@HuggingPapers Links are dead

Available in RL-tuned and fast 4-step Lens-Turbo variants: huggingface.co/microsoft/Lens huggingface.co/microsoft/Lens…

Microsoft just released Lens on Hugging Face A 3.8B parameter text-to-image model delivering efficient training and high-resolution generation up to 1440×1440.

@akari_w0r1d Ah I’ve been waiting for a head to head re Midjourney!

Krea 2 での画像生成、衝撃的にいいかも！Midjourney よりもいいかもしれないと思ったツールは初めて！

@krea_ai I missed these I guess - I was hoping to review the new model in my creative Ai newsletter out in a day or so. it’s legit - arnicas.substack.com

sharing more Krea 2 access codes! K2-NF55KA / K2 TCA7ZX / TMYVGV – each one will work for 80 users. link for redeeming them below 👇

@Miles_Brundage So sorry.

RIP my beautiful boy Tippy. x.com/lfschiavo/stat…

@HBCoop_ How does it compare to Midjourney? Weird I’m struggling to find anyone mentioning that

Had the pleasure of testing this out over the past week. Very, very good model. Quality aesthetics, fast generations along with Moodboards and Style transfer. Excited to explore more with Krea 2 ↓

@enjalot Actually looks a little useful…

very weird bug but fun effect

@ComfyUI @NousResearch Will that work with your cloud?

When your tool is open source and free, your creativity has no ceiling. The ComfyUI skill in @NousResearch Hermes Agent lets you compose sophisticated workflows by chatting to an agent. Try it today.

The most female-led product org in tech right now: Chief Product Officer: Ami Vora Claude Code/Cowork Head of Product: Cat Wu Claude Code/Cowork Head of Eng: Fiona Fung Claude Platform Head of Product: Angela Jiang Claude Platform Head of Eng: Katelyn Lesse Research Head of Product: Dianne Penn President: Daniela Amodei (Also, the fastest-growing company in history)

@fabianstelzer I’m interested … how did you run it? What harness?

I don't really ever trust benchmarks, so ocassionally I'm stress-vibe-testing a bunch of new models on some very complex agent work (hundreds of tools, not your simple coding agent stuff) and to my surprise, the clear winner is DeepSeek V4 Pro ...which is also the cheapest!