Lica came out of an obsession: to give everyday people access to powerful design agents to tell their stories, without the tool itself becoming an inhibitor to creative expression.
As we went deeper, we kept running into fundamental problems that remain unsolved at the frontier and stand in the way of everyday people using these models to truly express themselves. Easy to one shot a stunning image; precisely shaping an output to match what someone has in mind, and more importantly, understanding their version of taste, is much harder.
We spent the last three years chasing many of these problems: layer level editability, quantifying taste, building RL loops that could hill-climb on specific creative tasks, and systems that intelligently allocate tokens across models while preserving creative intent. Along the way, we built the data, evaluation, and model infra to turn that research into products used by some of the world's largest consumer brands across fashion and CPG.
What @thisisgrantlee and the team have built at Gamma is special. Serving 100M+ users everyday, they have a unique vantage point into what still stands between frontier models and everyday adoption. And we have a shared commitment to solving those fundamentals. I'm incredibly excited to join forces and push that last mile forward.
Thanks to our early customers, employees and investors. Thank you for believing in us. Your conviction gave us the chance to pursue an idea we deeply cared about. We uncovered answers to many of the questions we set out to solve, and I’m excited to see them reach millions in this next chapter.
Shoutout to @amasad and @pirroh for being day 0 believers in me. I vibe coded and spun up agents in @Replit back in early 2023 before it was a thing, and a lot of what @world_lica became later wouldn’t have been possible without you both.
Thanks @spc for nurturing one of the best communities any founder could ask for. (no surprise that @GammaApp came out of it too)
Thanks to our investors @Accel , @spc, @villageglobal , @spc , @amasad, @pirroh , @snsf
Shoutout to @adityaag, @ditzikow, @SaraIttelson, @skg_accel , Amrit, @bencasnocha, Rei-Ling. You guys were always generous with your time through all the ups and downs.
I’ll write more about the journey in the next few weeks. Meanwhile, if you’re curious about our new research direction, send me a DM!
Thanks @IndianIdle for covering this.
Can you train AI to have taste? We think it's worth finding out.
I’m thrilled to announce our acquisition of Lica.
@pritopian and @purvanshi_mehta share our belief that visual communication should be easy and accessible. Together, we plan to build a new research team exploring what comes next:
How do design, layout, and visuals shape attention?
Can qualities like taste and trust become learnable signals?
And how can we help our customers make their ideas matter?
Visual intelligence remains remarkably jagged at the frontier. We’re investing in the research and talent needed to help AI truly understand design, taste, and communication. Our ambition is to create new visual formats that surprise and delight without becoming harder to use.
The team has already hit the ground running, and we can’t wait to share more in the coming weeks.
Since I started using Dots, I haven’t opened my other phone apps. I just dump my thoughts on the go, and Dots is a memory bank + personal assistant that can remember everything and run the show. Reply to emails, make calendar invites for follow ups and generally just get stuff done. But, haven’t fallen in love with voice calling feature yet.
JD Vance talks about how his Indian wife and her family taught him how to evolve past his hillbilly ways, and how she once prevented him from beating up a motorist.
Why not support more immigrants to teach poor Americans how to behave? He now considers them a cultural threat.
My top muse usecase is to send it a reel where a creator says “comment DM and I’ll send you these products” , and let muse find them for me. I do wonder when they’ll share more about the second order effects of muse on other product surfaces like reels, marketplace. Who are the influencers curating products for? If muse is really the one devouring their content.
Some stats on Muse from Meta's earnings call:
• 10M+ people in the US and Canada set it up in ~6 weeks
• ~25% of users use it on any given day
• Shopping is the top use case, used by 25%+ of weekly actives
• Average purchase size is ~$200
@gavinpurcell Agreee! Im sure they can mine some gems from https://t.co/zUIHHZdzfY when it was super active with fans uploading spin off manuscripts back in the day.
@TaliaGold I think I was pretty convinced about Tesla’s safety engineering after reading about this family. They survived a 250 ft fall off a cliff. No serious physical injuries https://t.co/Nv5z9CeBat
@jennyAI I’m wondering what’s the strategy here ? Do they expect only low performers to be the ones taking the option for a voluntary exit? And is that a more lucrative deal than a severance?
there’s this vintage short essay 'i want a wife' by judy brady, a tired 1970's housewife who asks for a wife of her own to offload tedious tasks to
and it is also inadvertently one of the best product roadmaps i've seen for personal ai assistants/ home robotics
@SoniaBaschez@Waymo lol you’re right. I guess predictable revenue from loyal customers might still be lower than on demand riders willing so pay for surge pricing? But who’s waiting 25 mins+ and paying $52 for 4 miles at 9:30 AM in SF. 😭
Is there no way to reserve daily @Waymo rides beforehand during morning commute hours and not pay exorbitant prices 🥲the wait times are 20m+ and I’m forced to take an uber.