We’re excited to introduce Taste Labs.
Our mission is to end AI slop. We’re building the data and infrastructure layer to give AI models and agents taste.
And today we’re coming out of stealth, announcing our $18.5M seed funding, co-led by @CRV and @AmplifyPartners
AI has nailed objective domains and made it easy to generate anything. But it still feels off. Now, the challenge is judgement. What fits, what feels like you, what’s GREAT. This requires turning a fuzzy, subjective domain into something we can measure and codify. We’re starting with design.
There are two sides to cracking this, the foundation model layer and the agent layer:
- We’ve already been working with the top frontier labs to evaluate and improve their models, crafting the right post-training data and RL environments.
- We’ve also been working with app-layer companies to build the context and verification tools for their agents to produce better, more on-brand, more creative outputs.
We want a future where AI feels right.
If you’re passionate about this mission, join us!
lots of tiny widdle babies in the comments. you awent wich enough to affowd a 1% tax? you awent wich enough to live in NYC full time? so sowwy maybe one day you will be big boy wich but i guess not now 🥺
Intelligence too cheap to meter: This 10 minute clip took 8 hours to creaste and cost around $60. That's fast and inexpensive for an excellent anime clip.
Soon everyone will be a movie-director.
And I think many still don't understand what that means. We've crossed the threshold where large teams are needed for video production. Even SeeDance's clips are the worst they'll ever be. From here on out, it only gets better.
How Learning Something New Merely For The Joy Of It Taught Me How To 10x My Productivity
My Agent Read All The Booker Prize Winners So I Don’t Have To Waste Time On Such Frivolous Pursuits
If You Don't Know Zen Koan About The Infinite Capacity Of Love, You Will Be Replaced By
Today, we are unleashing https://t.co/AryN8nzbCt
This platform puts the worlds most powerful frontier AI models directly into the hands of every American warrior.
We will continue to aggressively field the world’s best technology to make our fighting force more lethal than ever before.
And all, American made.
What if Harry had just wished Voldemort away? What if Aladdin had to become a prince without the Genie?
the AI-EdTech industry is using lamps where we should be using wands. That has consequences. Check out my piece here: https://t.co/HRxQ5OZyn5
this is not an upscale. this is "imagine what this pic would look like if it was better resolution". and the model does a good job *imagining*. but it *imagined* stubble, freckles, shirt pattern, glasses texture, and (if you look closely) his bottom lip.
training on synthetic data produced by gpt will result in a preference towards similar outputs. similar to how llm evaluators will prefer their own outputs even when authorship is hidden. it's all encoded in the weights!
I built this months ago then realized the winner was GPT every single time.
Even after plugging in Gemini 3.0, the winner was GPT 5.1, every single time.
Even funnier, if you TELL other models (Claude, Gemini, Grok) that the answer they are reading came from GPT (un-anonymize), they fold immediately and start correcting themselves based on GPT’s output.
This is why OpenAI is not finished yet.
another great session of the @ekkolapto x @medialab computational philosophy club. @will_hahn gave an excellent talk. the irony of tweeting this - social media is a perfect vector for memetic prions 😵
had a great time at this hackathon hosted at @medialab and created by @ekkolapto . here's my talk about a novel approach to understanding the nature of #consciousness (inspired by @Plinz talk early in the night).
what do you think?
https://t.co/rLHrIW0XL2
After a year of work, @CitytutorDC has uncovered the principles of what we’re calling “the connected forest” strategy.
Read their white paper, Establishing Roots: Implementing Citywide High-Impact Tutoring in DC: https://t.co/rkZQPpaRAu