Jason Ma ✓

@JasonMa2020 · joined 26 Aug 2018

Co-founder @DynaRobotics Prev: @GoogleDeepMind, @NVIDIAAI, @Penn, @Harvard.

124 195Followers

998Following

1 053Posts total

116.1KViews on collected posts

Posts recentes

Jason Ma ✓@JasonMa202027 Aug

This is our release that I am actually most excited about because it shows and proves what truly matters. Deployment is the ultimate prize and the only reliable eval for robotics. After having worked on so many research projects and models in my career, where I saw so many new 11K views · 117 likes · 15 reposts · 10 replies Open on X →

Jason Ma ✓@JasonMa202018 Aug

I will be at #Actuate26 giving a talk on the model and infrastructure behind dyna-2. Excited to chat with everyone about scaling robot foundation models and deploying them in the real world! Please reach out if you'd like to chat! https://t.co/PnwKh6hifZ 11.9K views · 90 likes · 13 reposts · 7 replies Open on X →

Jason Ma ✓@JasonMa202017 Aug

At million-hour scale, the bottleneck is not the robots or the GPUs. It's everything between the robots and GPUs. Our cracked infra team did an amazing job scaling up and revamping our ML infra to enable the dyna-2 breakthroughs! Read our new infra blog to catch a glimpse of the 20.9K views · 147 likes · 10 reposts · 12 replies Open on X →

Jason Ma ✓@JasonMa202014 Aug

predicting the future is key for unlocking generalization, great work by Jesse and co! 9.4K views · 33 likes · 2 reposts · 0 replies Open on X →

Jason Ma ✓@JasonMa202012 Aug

Robot WAMs typically predict future RGB latents trained for pixel reconstruction...but why stop there? We propose Flex-π, a WAM that jointly denoises multiple visual streams jointly, enabling the WAM to inherit strong 3D, object-centric, AND spatio-temporal priors for action 26.2K views · 160 likes · 23 reposts · 5 replies Open on X →

Jason Ma ✓@JasonMa202010 Aug

Super excited to share Dyna-2, the first robot foundation model trained on over 1 million hours of human data. At this scale, we saw the emergence of a cross-embodiment transfer scaling law: training on increasing amount of human data not only improves model prediction on https:/ 36.7K views · 298 likes · 30 reposts · 22 replies Open on X →

Em comparação com contas do mesmo porte

6 posts dos últimos 90 dias, ao lado da faixa de 100K–1M seguidores. aparece para muita gente, mas poucos desses espectadores reagem.

Mediana de visualizações16 436esta conta4 600mediana para 100K–1M

Alcance, %13.23%esta conta1.64%mediana para 100K–1M

Engajamento, %0.88%esta conta1.25%mediana para 100K–1M

Métrica	Esta conta	Mediana para 100K–1M	Proporção
Mediana de visualizações por post	16 436	4 600	3.57×
Alcance (visualizações ÷ seguidores)	13.23%	1.64%	8.07×
Taxa de engajamento	0.88%	1.25%	0.70×

Outras contas desta faixa → Comparar com outra conta → Como estas referências são construídas →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

36.7K10 Aug

26.2K12 Aug

9.4K14 Aug

20.9K17 Aug

11.9K18 Aug

11K27 Aug

Last 6 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.98%10 Aug

0.72%12 Aug

0.37%14 Aug

0.82%17 Aug

0.94%18 Aug

1.30%27 Aug

Reactions — likes, reposts, replies and quotes — divided by views. Median for 100K–1M accounts is 1.25%.

What the audience does

Likes73.1%845 in total

Reposts8.0%93 in total

Replies4.8%56 in total

Quotes1.3%15 in total

Bookmarks12.7%147 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.