TLDR
EA is a community where time tracking is already very common and yet most people I talk to don't because
- It's too much work (when using toggl, clockify, ...)
- It's not accurate enough (when using RescueTime, rize, ...)
I built https://donethat.ai that solves both of these with AI as part of AIM's Founding to Give program. Give it a try (and use discount code "EA" after the 14d trial to get another month free).
You should probably track your time
I'd argue that for most people, your time is your most valuable resource.[1] Even though your day has 24 hours, eight of those are already used up for sleep, another eight probably for social life, gym, food prep and eating, life admin, commute, leaving max eight hours to have impact.
Oliver Burkeman argues in his recent book Meditations for Mortals that eight is still too high - most high impact work gets done in four hours every day - the rest is just fluff and feeling busy.[2]
Now, how do you spend those four hours? When it comes to our other scarce resource - money - most people and companies keep budgets, there is a whole discipline of accounting to make sure it's spent wisely. But somehow, for time, we just eyeball it.
When tracking time, the objective isn't to set a number and play "number go up." The objective is to understand where you spend your time and help you prioritize and plan better. AI is estimated to increase workforce productivity by 5%.[3] Imagine the increase of productivity if everybody would be better at planning and prioritization.
One last reason that is often overlooked: Tracking time can reduce anxiety and guilt. We often feel like we should "do more" but there is always more to do. By setting realistic time-based goals like "work 4h/d on project X" we have a clear measure when we achieved the goal and have also full control over the outcome.
If you want to dive deeper than just these handwavy arguments into why it's useful, check out the LW post by Lynette, or the discussion in this post on how much time people work.
It just got easier
I talked to a lot of EAs this year at various EAGs and EAGxs and the most common reason for why they are not time tracking is "I tried it but couldn't build the habit". When poking a bit more that was usually because it's too much work to manually track and when doing it automatically the data quality wasn't high enough to keep doing it.
Earlier this year I joined AIM's Founding to Give[4] program and built donethat.ai - a time tracker that uses AI to fully automate data capture and analysis. It literally takes two minutes to set up and you can forget about it afterwards.
It's not perfect yet, it never will be, but it already helped a bunch of EAs be more productive (and they helped me to improve the tool, thank you!!!). If you feel something is missing, always let me know, and feel free to check out my list of similar tools - maybe you'll find something that works better for you.
If you want to give it a try, please check it out. It's free for 14 days and you can use discount code "EA" to get an extra month for free after that. If you're thinking about using this as a team, please reach out, I'd love to work more with teams as I think that's where the biggest impact is but it's also more sensitive than when just tracking individually.
I've been using DoneThat for a few months and it was amazing to see an accurate picture of how many hours I work and what I did, and how often I switch between tasks. All that without having to track time! I realised I work too many hours so it helped me to work more sustainably. Also really nice to see how long each task took, sometimes I think "oh I think that was an two-hour task" and then I look at DoneThat and it shows as a 3 or 4 hour task because of other smaller sessions, which is useful to know.
Thank you for all the feedback along the way!!
I used DoneThat for a while and also highly recommend it, especially given the low cost (5$/month)
As a piece of feedback, I think you should have included this video in the post: https://www.loom.com/share/53d45343051846ca8328ccd91fa4c3a8 and people should look at it before deciding whether to download it. It made me feel much more confident in the privacy aspects (especially when using one's own Gemini API key)
If you upload it to YouTube you can also easily embed it in a bunch of places (including this forum)
Thanks!! It's embedded on the landing page but fair point, would've probably helped to also embed it here!
How risky is it if I forget that the app is taking screenshots and I enter my password into some form (where the password is visible), or just have open a document containing sensitive information? Is the bad thing that happens that the app might have a summary like "The user enters password xyz into this website" or "The user opend a document containing xyz"? And then I just have to either find the summaries from that screenshot in the calendar and delete them or hope that no one who follows me would see that specific summary?
I feel like having something screenshot my computer every 5 minutes and write a summary is something a IT security person would warn me against. (I don't know much about computer security, so maybe this fear is misplaced.)
Hey Karl. Yes, a IT security person should warn against this. That's why I have taken lots of measures to keep this secure, most importantly, never actually storing the raw data and prompting AI to never extract sensitive information (like passwords). See the privacy section on the landing page https://donethat.ai Right now also working on another feature that users can toggle to auto-delete the minute-by-minute data (and only keep the high level summaries) at the end of their day for even more privacy.
I'd say do mention this when applying for early stage orgs (e.g. AIM or AIM incubated charities) and they really value this attitude. For orgs which don't have a budget or have a very small budget, it would make a huge difference to hire someone who is good and doesn't cost too much. Also it may mean that in some cases you will be hired over some people simply because you are the best in that range of pay (assuming you are).
I'd say that for bigger, more established orgs it shouldn't make a difference, as 30-50k probably don't matter to them as much as hiring the very best person, especially for harder to hire roles, so they are more likely to choose someone a bit better than you even if they have to pay them a fully salary. I probably won't mention this until the interview, where I'd ask them what salary sacrifice or payroll donation options they have because that's something that you'd like to do.
Also should you get a job where you end up getting paid a lot less than everyone else, ideally give your employer plenty of notice (e.g. 3-6 months), because once they start relying on you, it may take them ages to not just find someone to replace you but to also fundraise for the normal salary, assuming they don't have it in the budget. And for a small org it can be very stressful. Sure it's the leadership responsibility to account for that ideally, but it can be very helpful to plan, especially considering some smaller non-profits don't have long runways.
On a personal note, I'm unsure what your financial situation is, but I'd also consider making sure that you have enough savings to last you should you lose your impactful job (e.g. 12-24 months of expenses), as it can take ages to find a new one even if you have more experience. You can also use this time to volunteer as I personally found it invaluable for my org when people who are well-off financially could volunteer 20-40h. It made a huge difference to us in the first few years.
I like the idea but do not but am a little fishy on the security part (e.g. trusting remote LM providers). If I have a local server running a VLM, is there a way to set the app to only communicate with that server?
(Not knocking on the idea / executtion! I think this will work for a lot of people, e.g. those who are down for OA browser, etc)
Was planning to have this eventually but so far nobody asked. Happy to figure this out together. I'll DM you
It would be amazing if one could use it with a local LM in LMStudio
Cool, thanks for sharing!
I currently use Timing.app, and have been recommending it to people. Is donethat different in any ways? (TBC, "it has all the same features but also supports an E2G effort" would be sufficient reason for me to consider switching).
Hey cb, thanks for sharing, will add it to my alternatives list. From what it looks like timing.app is more mature but uses a different approach that could lead to less accurate results. To me it looks like it mainly uses active windows (like you're on chrome, on slack, etc.) while DoneThat analyzes your actual screen and thereby gets more detailed data for better understanding. Would be curious if you'd run them side by side how it works for you.
DoneThat is also significantly cheaper (at least for now) and Christoph is very responsive to feedback/requests (once replied to an email within 6 minutes)
It says that it is downloadable on Linux, but I just started using it. Are there clearer instructions on what buttons to press/what words to put into the terminal?
When you hit download you should've seen a popup with instructions. But basically just execute it with the parameter --no-sandbox. You have to set it up for autostart yourself, let me know if you need help with that
I don't know how to use a parameter on a file, or how to execute it. How do I do that?
Looking forward to trying this! I've had RescueTime installed for a while, but this week looked at the analytics & didn't find them too helpful. I'll probably use DoneThat in combination with a desktop/mobile-hybrid while DoneThat is desktop-only.
Let me know how it goes! Thinking of mobile but iOS is pretty restrictive with the data you can record (probably a good thing)
I use https://dailytimetracking.com/ and find it to be fairly unburdensome, though it might not be great for those with more porous work-life boundaries.
Oh nice, I think that's one of the leanest manual trackers I've seen. Will add it to my list of tools in that space.
Btw for anyone this helps: My Norton antivirus did not like the download. I decided this was high trust enough that I disabled it and as far as I know nothing bad happened. I could turn it on again after installing the excellent software.
Ah yes that unfortunately happens sometimes. Because the software offers (optional) keystroke tracking it has some dependencies that I'd imagine trigger antivirus software.
Yesssss!!!! I am trying it right away. I also think for many here, timing is important to set limits. Like cap your work week at 50 or at most 60 hours (or less if you have caretaking responsibilities). That way you don't let guilt push you into unhealthy territory. That's how I use timers. Also great for parents that are both ambitious to make sure one does not get a career advantage by feeling more nervous or something.
Yes fully agree that capping is important. I'd probably cap it much lower (I guess I average about 20-30h/week of actual work on DoneThat). I like this post where people share how many hours they work https://forum.effectivealtruism.org/posts/byMQvEHWur23bLpQw/how-much-do-you-actually-work#GBXjoJZudHpLh72Mg. Anecdotally I also talked with somebody who tracked productive hours in a high-paid US tech job, averaged about 4h/d and got promoted with that.
I use DoneThat and like it, thanks for building it!
This looks super cool! But is it only for linux for now? I also didn't realize how much the AIM association would pique my interest, reading that is what made me actually open the website! Either way, looks great, thanks for sharing!
It supports linux, mac, and windows. It auto-detects the system based on your browser... maybe that went wrong. You can select the platform here: https://donethat.ai/download
Is it possible to use it on Chrome OS somehow? It auto-detects that as Linux but I think it won't work if I use the Linux installer. I'm pretty sure it would be installable as a browser add-on but then not sure if it would work when you're using other programs.
I don't have one so we'd have to try this together but according to ChatGPT you can activate Linux in your Chrome OS. Open Settings → Developers. Under “Linux development environment (Beta)”, click Turn on. Then you should get a terminal. And from the terminal you should be able to execute the app (./path-to-file --no-sandbox). Might need to install dependencies before... if you want give it a try and DM me - we can do a video call and see if we get it working together
Can I take you up on the offer to do a video call and see if we can install it on Chrome OS? Will DM you
DMed you
LMK what the outcome is (if it works smoothly on chromebook after you do this stuff)!
Just did a few tests and looks like it's impossible to get to work on chromebook because the linux instance is sandboxed and doesn't have access to the chromebook screen for screenshots. So doesn't work on chromebook unfortunately!
I got to the terminal but wasn't able to access the download and gave up at that step because for some reason I assumed it would only install the app for the linux development environment as opposed to the rest of Chrome OS. I'll try again, and email you if I can't get it working.
I use Toggl for tracking how much time I spend on all work tasks and I highly recommend it. It has both a website and mobile app which I find intuitive. Although using it requires non-zero effort, I find that using Toggl helps me start tasks and feel more intentional and aware about how I spend my time.