$FIG Q3 results:
→ $391M in revenue, 43% YoY growth rate and $16M above high-end of prior guidance
→ 131% net dollar retention rate
→ 36% YoY growth in non-GAAP gross profit
→ $1.7B in cash, cash equivalents and marketable securities
→ Raising our outlook for FY 2026
Excited to share more about the work behind Figma's first-party model: a specialized model for design editing that competes with top models.
When we set out to build an agent, we wanted a model that understands design concepts and Figma's tools to help you quickly make edits on the canvas.
We started with supervised fine-tuning (SFT) using Figma's native file format, pairing designs with synthetic edit requests. Deterministic validators, visual judges, and a self-healing pipeline helped evaluate and refine our outputs. We've since added reinforcement learning (RL) to our recipe to push for more improvements.
Design is subjective, but we can still evaluate structural correctness, visual alignment, and request fulfillment using judges calibrated with expert designers. That includes checking whether an edit preserves everything it should. AI models often create unintended changes like dropped text or mismatched chart labels that we need to catch.
Proud of the work of our incredible AI research team. Read more about the model here, and if this work sounds interesting, come join the team!
Do you even use Facebook? Let's do a real world test to see what goes on there. I opened up my Facebook feed, and categorized the first 50 posts that I saw as either real content posted by my friends I follow, suggested content, advertisements, and attempts to get me to interact with "people you may know."
Of those 50 posts, only 15 were real. 26 were suggested. 7 were ads (although many of the suggested content was secretly spam clickbait). There were 2 attempts to expand my friend list.
This is embarrassing. Only 30% of the content is from actual connections. 70% is engagement farming and advertisements. 70%, Yann.
Here's some examples of the weird suggested content I received:
1. A picture of Keanu Reeves holding up a t-shirt that says "If the government says you don't need a gun, you need a gun" from the "Native American Community."
2. A Far Side Comic about border police stopping a car and saying "Excuse me, sir, but could your entire family please step out of the car? ... Your faces are not in order." This is supposedly from 1992 but who knows if it is real.
3. Some bizarre comic about anxiety, and at the bottom it says "NSFW, Adult Only Comics, PinUps and Animated Gifs" with a link to some random Patreon account.
4. This one is something posted from an actual friend. It's a meme of Donald Trump with the phrase "we should worry about death camps if he wins the election." There is no attempt to provide fact checking, despite it just being a large reshared meme account.
So let's review the current state of the product you clearly don't use. At least 70% of it is spam, clickbait, potentially scams, and misinformation run amok.
So if I'm clearly wrong, what website was I just on?
Apparently, you need 1-2 publications at top-tier ML conferences to get into a top school for a PhD: https://t.co/EVOn6V1XWw
I wonder if getting into top schools is still worth it, then?
1/5
Us: You gonna keep that lightning port forever?
Apple: Don't make me
Us: Maybe it's time to move on
Apple: You don't know what you're doing
Us: Seriously USB-C is the future
Apple: I'll do it
Us: Wait what
Apple:
Excited to share that my paper titled "Towards better confidence estimation for neural models" has been accepted at ICASSP 2019. Thank you @icassp2019 !
Let's set a world record together and get the most retweeted tweet on Twitter. Beating the current world record held by Yusaku Maezawa @yousuck2020 (5.3 million!)
We got this 🙌
Photo of one single atom (the bright dot at the center of the pictures). 🧐It's being illuminated by a laser and held motionless thanks to an electric field. It absorbs and re-emits light particles sufficiently quickly for an ordinary camera to capture it.