Product & growth person. Tinkering. Recently recommendations and growth product @Instagram. Previously @Dropbox self-serve growth, @McKinsey @pennmandt.

Joined December 2008
Congrats on the launch of @genmoai - @ajayj_ and @parasjain! It's addicting to play around with 😉
Announcing Genmo Video, a generative media platform with a new text-to-video model that can generate immersive live artwork from any prompt or any image. What will you create? 🎨▶️ Free public access: genmo.ai Discord: discord.com/invite/u7SRpXHhp… 👇1/n
1
6
857
Aditi Jain retweeted
Releasing Skyplane, a new open-source tool to move huge datasets between clouds. Skyplane is: 1. 🔥 Blazing fast (110x faster) 2. 🤑 Cheap (4x cheaper) 3. 🌐 Universal (AWS, Azure and GCP) Read more: medium.com/@paras_jain/skypl… github.com/skyplane-project/… skyplane.org 1/
6
59
14
247
Excited to see @ajayj_ and team's work come out. Have watched the early versions of my brother @ajayj_'s 3D generation work going back to Spring 2021, and image generation going back to 2019, and the 3D generation is getting good 🔥 Can't wait to see what's next.
Happy to announce DreamFusion, our new method for Text-to-3D! dreamfusion3d.github.io We optimize a NeRF from scratch using a pretrained text-to-image diffusion model. No 3D data needed! Joint work w/ the incredible team of @BenMildenhall @ajayj_ @jon_barron #dreamfusion
1
4
Excited for your talk @parasjain!
I am very excited that our work applying approximate computing to the Google TPU won the best paper at MLArchSys@ISCA2021! Our talk is today at 10:15am PT at #ISCA sites.google.com/view/mlarch…
2
What if your IDE crashed with specific variable names? 😱 #ML4Code is powerful, but models like BERT are sensitive to adversarial code edits, performing worse than random guessing! We find that contrastive learning is more robust. Check out ContraCode: arxiv.org/abs/2007.04973 👇
2
At Dropbox one time I manually covered the candy jar near my desk with printer paper — it lasted 3 days and confused the hell out of the staff who refilled the snacks. Nice to see actual experimental validation though.
Google's NYC office ran an experiment: - Version A: put M&Ms out in the open - Version B: put M&Ms in opaque containers People ate 3 million! fewer calories in a single month in version B Tiny friction goes a long way. To create healthy behavior, start with healthy defaults
2
Saw this Barbie that teaches meditation, in a holiday gift guide. Cool to see meditation starting to become more mainstream and efforts starting to target kids. target.com/p/barbie-breathe-…
1
2
To contextualize the size of these numbers, I found that est. global VC investment in 2019 was $295B. Thus, these 5 companies' investment in R&D and CapEx in 2019 was equal to 63% of global VC investment in 2019. Truly astounding. Source for VC #: news.crunchbase.com/news/the…
These five companies invested nearly $200B in R&D and CapEx in 2019: Amazon $53B Google $50B Microsoft $29B Facebook $29B Apple $26B The scale at which these companies are investing is astounding.
10
Holiday season: the time when you do all the administrative tasks you never have time for year-round. Seeing the optometrist, going to the DMV, reflecting on the year, expressing love to your family. #joke
1
4
Proud of my brother for leading work on this with a team at Uber ATG's research group in Toronto over the past 1+ year!
How do drivers predict future pedestrian behavior? Our map-aware model, the Discrete Residual Flow Network, predicts multimodal and uncertain pedestrian behaviors over long time horizons. We'll be giving a talk at #CoRL2019! Paper: arxiv.org/abs/1910.08041
2
Aditi Jain retweeted
How do drivers predict future pedestrian behavior? Our map-aware model, the Discrete Residual Flow Network, predicts multimodal and uncertain pedestrian behaviors over long time horizons. We'll be giving a talk at #CoRL2019! Paper: arxiv.org/abs/1910.08041
8
2
15
Interesting example of a marketing push leading to a cultural "fact" - "10k steps a day" was never grounded in science, it's more like 4k-7.5k a day: npr.org/sections/health-shot…
1
4
Really well said! Executing with 100% commitment removes existential dread, so teams can focus and move fast. But keeping a healthy amount of existential dread ensures you can change directions nimbly based on a constant stream of new inputs. The only constant is change.
Under appreciated skill: ability to think probabilistically, but execute with 100% commitment to the decision. Always impressed with people who can seamlessly shift between those two gears.
1
2
Having principles can help navigate these decisions as well as realizing there is no perfect. But I haven't yet fully internalized and lived by this. Would love to hear how others think about this. Note: I realize this is extremely privileged and I try to be grateful everyday.
1
1
Where I'm at currently is: do something, do something good, do it with no hesitations, do it with people you enjoy working with & learn from.
2
This x100. There is no playbook for employees when joining a fast growing company (moreso at a startup). Read this as "how to have the most impact" in a fast changing, uncertain environment. There should be a followup thread re: how to be an effective mgr at an early/growth co.
1/ How to get promoted at a startup: At larger companies career tracks are (too) well defined, and there are clear paths on how to move up. But at startups there isn’t a playbook for how to get promoted.
7
It's my job to show others who I truly am, and to develop the resilience to be okay when they're not ready to accept me or my ideas.
1
9
Now you can track your metabolism with lumen.me/, fix your diet with @getonefix, and get swole with @tonal. Interesting week of shiny, new digital wellness product launches. Just costs $199, $99/mo, and $3k respectively ;)
7