Dan Grover
@DanGrover
Followers
12K
Following
230K
Media
734
Statuses
14K
Product @standardbots. Prev. Meta, Tencent.
Oakland, CA
Joined February 2007
Figma needs, like, a way to quickly add notes and musings alongside groups of mocks (when you're doing explorations). Comments isn't quite right -- I use text boxes but there's friction in having to style the text and discard whatever formatting was applied by default.
3
0
1
I would consider myself a tech bro and have listened to the usual tech bro podcasts since podcasts were invented. Somehow, now, in AI age, they make me kind of nervous and feel more like work instead of fun. Maybe I need more true crime.
4
1
18
RT @Russell_MFG_USA: Just dictated 9 pages of single spaced text to describe picking up a steel bar, loading it to a saw then loading cut….
0
4
0
Swapped out the old robot in my WeWork cube for the new model RO2 my company makes! Exciting to finally upgrade!
3
0
17
Novel hardware design choices aside, I feel like more robots should be wearing straw-boaters.
Your bimanual manipulators might need a Robot Neck 🤖🦒. Introducing Vision in Action: Learning Active Perception from Human Demonstrations. ViA learns task-specific, active perceptual strategies—such as searching, tracking, and focusing—directly from human demos, enabling robust
3
1
10
It's like the "Pray tell, Mr. Babbage. " anecdote. I feel like I'm on crazy pills.
1
0
15
I feel utterly confused when people speak of generating PRDs with a prompt. If you don't know the requirements, how would the AI? If the prompt is meaningful, then just make the prompt the PRD.
Prompting is the new wireframing. The most valuable PMs in 2030 won't be those who write the best PRDs, they'll be those who can "talk" to AI most effectively. Start treating prompts like code: precise, testable, version-controlled.
51
13
381
I wonder if the danger with people relying on AI for personal problems isn’t that they’re sycophantic or randomly tell you unconstructive things but that people shouldn’t have *that* much therapy. Like there should be a bit more friction.
0
0
6
I have not been monitoring the situation at *all*. What is a good place to start for people who are new to situations?.
4
0
20
I feel like a good gimmick for some work of epistolary fiction would be a novel in the form of the protagonist’s conversations with AI. Has anyone prominent done this?.
If you ever wanted to stare directly into the Facebook Boomer reactor core, now you can
2
0
3
I switched to Linux as my daily driver a couple years ago and have been annoyed at its inability to match the overall level of design polish on the Mac. With these latest announcements from Apple, Linux has finally caught up!.
0
0
11
Love everything about this. More knives, less avocado slicers, please!.
New longform @inkandswitch essay! 📜. Malleable Software.by me, Josh Horowitz, @pvh and Todd Matthews. It's about why people need agency over their software tools, and how to make that happen. Here's the quick tl;dr. 1/.
0
0
10
Thing with places where you can't talk to users is's usually a separate, priestly class of UXR/designer/ops that make a big deal out of abstracting into personas and stories, etc. Sometimes helpful, sometimes divorces us from intuitive signal we get from talking to actual humans.
my friend & i were discussing how we both worked at places where the people working on the products just simply stopped talking to their users. this may seem obvious to some but why is that? it’s because incumbents have two customers: shareholders & users. the longer they have.
2
0
10
Part of me always wonders whether "robot GPT" is wrong goal. In short term, particularly in industrial env, there are oodles of un-automated tasks that transformer models could be applied to. The robot doesn't need to generalize or follow natural language prompts: it needs to do.
Assuming that we need ~2 trillion tokens to get to a robot GPT, how can we get there? I went through a few scenarios looking at how we can combine simulation data, human video data, and looking at the size of existing robot fleets. Some assumptions:.- We probably need some real
4
1
22
Sometimes I am glad I moved to Oakland in my late 20s. I get to just run around the lake and eat hot chicken sandwiches. That is pretty much what is going on over here.
pov: living in SF in your 20s–30s is basically the Hunger Games:. - the reaping: founding engineers who give up on their founder dreams .- Blue Bottle SoMa is the pregame showcase.- everyone is trauma bonded, no one wins .- Katniss: emotionally unavailable tech bro/sis you keep.
1
0
9
It is funny sometimes in a startup watching the handful of ex-FAANG people trying to replicate certain rituals that served a purpose in that system, like they're putting on a little pageant.
85
90
4K
What if you are a secret third thing, a PM that can code?.
Called shot:. AI will help software engineers become product managers faster than product managers becoming software engineers.
7
1
61