To the best of my knowledge, Waymo still has humans in the loop as Fleet Response agents that the vehicles can call for remote assistance when they aren't sure what to do. Caveat that the number needed likely isn't on the same order of magnitude as human drivers, but the job is likely higher paying. I could see a scenario where these should be locals for both latency (ChatGPT says SF to Miami RTT latency might be 80-100 ms and I don't believe the humans really teleoperate the vehicles, so that may not be meaningful, but that might be a bigger deal for international expansion) and knowledge of tricky intersections or road quirks in the city. They could also potentially help with labeling quirky city-specific scenarios and other various evals.
Love this idea! The Whiteboard Gym explainer video seemed really text-heavy (although I did learn enough to guess that that's because text likely beat drawing/adding an image for these abstract concepts for the GRPO agent). I found Shraman's personal story video much more engaging! https://x.com/ShramanKar/status/1955404430943326239
I've been trying to eliminate multi-tasking as much as I can, but the nature of startups day-to-day and even what seems like a single/monotask when zoomed out now often involves context switching (For say, investigating and fixing a user-reported bug, I might have to toggle between VSCode, localhost in browser + the DOM inspector or console, our bug tracker, our support ticketing tool, Slack, and sometimes the Cody window in VS Code/ChatGPT/Claude:
RT in pure trials: 448ms
RT in mixed trials: 710ms
Mixing cost: 262ms
RT in task-repeat trials (in mixed blocks): 710ms
RT in task-switching trials (in mixed blocks): 975ms
Task-switch cost: 265ms
RT in pure trials: 490ms
RT in mixed trials: 825ms
Mixing cost: 335ms
RT in task-repeat trials (in mixed blocks): 825ms
RT in task-switching trials (in mixed blocks): 969ms
Task-switch cost: 144ms
Second time (while listening to music --- I decided to do this since I've noticed it somehow decreases my latency in typing tests significantly):
RT in pure trials: 436ms
RT in mixed trials: 673ms
Mixing cost: 237ms
RT in task-repeat trials (in mixed blocks): 673ms
RT in task-switching trials (in mixed blocks): 746ms
Task-switch cost: 73ms
Edit: third time, also while listening to music:
RT in pure trials: 435ms
RT in mixed trials: 608ms
Mixing cost: 173ms
RT in task-repeat trials (in mixed blocks): 608ms
RT in task-switching trials (in mixed blocks): 700ms
Task-switch cost: 92ms
I suspect this "game" is also amenable to practice, and find it at least a bit weirdly addictive in the same way as Flappy Bird.
I don't think this game is all that representative of context-switching overhead, as my 4th attempt gives evidence that this improves quickly with practice:
RT in pure trials: 422ms
RT in mixed trials: 611ms
Mixing cost: 189ms
RT in task-repeat trials (in mixed blocks): 611ms
RT in task-switching trials (in mixed blocks): 602ms
Task-switch cost: -9ms
The "mixed trials" are naturally slower because I'm having to recognise 4 patterns instead of 2, but only by ~50%.
Whoa, a negative task-switch cost! I didn't take it multiple times, but it makes sense that having practice at this specific task probably improves both your overall response times and maybe more specifically improves the different trials.
What I'm curious about is whether we also get specifically good at say, task-switching between a code editor and say, Stack Overflow, over time
RT in pure trials: 463ms
RT in mixed trials: 833ms
Mixing cost: 369ms
RT in task-repeat trials (in mixed blocks): 833ms
RT in task-switching trials (in mixed blocks): 1040ms
Task-switch cost: 207ms
Awesome project! Reminds me of donothingfor2minutes.com from Calm, but with a different end goal of focus instead of calm.
Regarding mobile phones going to sleep, Wake lock [1] might help, unless you can reduce to 59s since I believe 1m is the threshold (make sure to request within the context of the user hitting "start"). Unfortunately on older mobile browsers [2], the best workaround I found was using this NoSleep library[3].
We thought about it but ultimately decided that owning a truck would detract from what we really needed to be doing which was getting truck owners to do the luggs
TL;DR: my co-founder and I hosted hundreds of 1-hour co-working sessions ourselves.
In the early days of what is now Flow Club, my co-founder and I had built several apps to try to help us stay in touch with busy 30+-year-old friends. It was tough to get any friends to even install the apps we made on Testflight, much less use them. They were busy with work or family (and the apps just weren't compelling enough).
We started asking friends to come work together on Zoom (during the pandemic) like we used to do at coffee shops. We wanted to add some structure to these, so we made them 1-hour co-working sprints with a screen-shared pomodoro clock and agenda (5 minutes to share goals, 50 minutes of working, 5 minutes to check in), sent out the Zoom links to friends, and then started pre-committing to times at the beginning of each week and sending that out to an email list. Within a month, we had hosted a couple hundred of these sessions between the two of us and couldn't keep up with the demand or requests for more times of day as it spread to friends of friends. When an early user who we didn't know IRL and then my partner each separately asked if they could also host sessions, we were blown away. We didn't think anyone else would want to volunteer to host. Then when we realized both of them were actually better "hosts" than we were, a lightbulb came on for us that we could stop doing the unscalable thing we had been doing and build for hosts.
>Working solo has its difficulties. For one, my income is somewhat tied to my productivity, and my productivity highly correlates to my state of mind.
Since going back to being primarily a maker after organizing my days around being a manager[1], and being an avid runner, I've redefined my relationship with "motivation" in a way that can be summed up succinctly by author Brad Stulberg: "You don't need to feel good to get going. You need to get going to feel good." I know that I am long-term very motivated, but day to day or hour to hour, "motivation" is a tricky word, because my energy and creativity waxes/wanes.
Agree with the author that structure is the most important thing for me to work around this. Even though makers dream of an open schedule, on the days where I'm off my usual routine, it's really tough to prioritize all the many things always on my plate. It's even tougher trying to decide to peel myself away from work to go for a run that I know will help me focus better after. Making the decision can be emotionally and mentally taxing, whereas if I rely on the default that I just go out for a run as soon as I wake up, the rest of the day just flows from that without the decision fatigue. Time-blocking or even just very simple structure like the OP has has been really effective for me. This includes a hard stop time each day even if it feels like I'm on a roll— my younger self would often borrow against my future energy, and that seemed to rarely work out in the medium-to-long term.
> This is combined with a lack of co-workers. Comrades in the trenches, if you will. And finally there's the ability to not do anything, which can be quite nebulous and dangerous if not managed.
For anyone who is a solo-creator struggling with this, "body-doubling" is a term from the ADHD/neurodivergent community that simply means "doing a task in the presence of another person". Surprisingly, they don't have to be working on the same task to help you feel like you have "comrades in the trenches". If you're interested, check out Flow Club in my bio.
You don’t think yourself in another emotional state, you act yourself into one (act as in you do stuff not act as in acting like an actor).
This is what I found combatting my social anxiety by approaching people on the street. Even after a decade of doing it, I am still as socially anxious as ever but simply giving some genuine compliments to a few fellow pedestrians loosens me up after half an hour (in Europe).
Good advice. Re: body doubling, I think I would benefit a lot from that, but I hate having a webcam turned on all day with strangers. Or even a microphone. I feel these body doubling apps focus too much on neurodivergent extroverts, and it's a damn shame.
Honestly, if I need strangers around, I'd rather work in a coffee shop, but ideally I just want something no more intrusive than an IRC chat to shoot the shit while the code is compiling.
Still crossing my fingers for body doubling that less intrusive.
Body doubling isn't only for neurodivergent extroverts. I'm an example of a neurodivergent introvert who needs body doubling to do more than 2-3 hours of work per day. I understand what you mean, though.
The problem with working from a coffee shop is that your strength and motivation to go there must come from you. You must get up, leave your apartment, go to the coffee shop, and decide to work there. For chronic procrastinators, it's rather tricky. A person like that needs enough motivation to do it but not enough to work alone.
Shameless plug - you can try us (https://workmode.net/). However silly it sounds, we provide body doubling as a service. Try the demo session - no registration is required, and it lasts from 15 minutes to several hours.
Some features that might convince you to finally give body doubling a shot:
- you connect with our employee (aka productivity partner), not a random stranger,
- you always connect with the same productivity partner,
- all sessions are 1-on-1. It's just you and your productivity partner,
- you don't have to enable your camera. We don't rely on you being our body double. We're happy as long as text chat/audio/screen sharing/webcam feed makes you productive,
- you can share your screen at a greatly reduced resolution. It's enough for us to distinguish between you watching Youtube/reading Hacker News and spending time in IDE/Excel/other work-related apps. We're unable to read any text on your screen, even at 32px font size,
- every day, we make sure that you connect with us. We make sure you tell us when you want to start the next day, and in case you don't show up, we'll call you and keep calling you.
Hmmm, what if you had a simple concept of pushing a button that's asking the other person, "hey, you still going?"
(I'm kind of reminded of video games where there are some "canned lines" that you can click on in the middle of the action and your character will "say" them to the opponent.)
Still going!
Still going?
10 mins more and let's break!
I think I'm done
Thanks for playing
Ok
Hmmm
Never!
Eh, not sure I want body doubling to peer pressure productivity.. mostly should be a way of very low key socialising and venting about code not compiling.
Oh interesting! I have a smart watch (albeit a Garmin instead of an Apple Watch) and purposely choose not to send notifications there because I'd look at it all the time. Do you find yourself glancing down at your watch even when you aren't getting notifications? Or do you have pretty tuned notification settings such that if you do receive one, it's likely important?
I've started leaving my phone in a different room, using it as a continuity camera, or literally tossing it on my chair or bed before sitting down to work, but inevitably I'll need it for 2FA via Google Authenticator— maybe it's time to move 2FA to another device, and once it's on my desk, it often gets picked up out of habit.
reply