4 ms·

I really feel this. I can make meaningful progress on half a dozen projects in the course of a day now but I end the day exhausted. I've had conversations with

by simonw 9mo ago

I really feel this. I can make meaningful progress on half a dozen projects in the course of a day now but I end the day exhausted. I've had conversations with people recently who are losing sleep because they're finding building yet another feature with "just one more prompt" irresistible. Decades of intuition about sustainable working practices just got disrupted. It's going to take a while and some discipline to find a good new balance.

bob_theslob646 9mo ago
Throw in the fact that clawdbot can work 24/7. It reminds me of why people wanted financial markets to be 24/7. We as a society should probably take a look at that otherwise it may lead to burnout in a not so small percentage of people
- billylo 9mo ago
  We should ask how the traders manage this. It's essentially 24/7 markets in the world. For them, the FOMO effects are even stronger... actual money earning opportunity.
- nextlevelwizard 9mo ago
  Why we as a society should give a fuck if someone can’t stop prompting? Unless you mean we as a society should make you pay for the damages your prompts are doing to nature?
  - bob_theslob646 9mo ago
    >Why we as a society should give a fuck if someone can’t stop prompting? It is not prompting, it is the constant feeling that you always have to be "on."
    - nextlevelwizard 9mo ago
      Still dont see how that is problem rest of us should worry. Instead of you just having ounce of self control
onlyrealcuzzo 9mo ago
> I've had conversations with people recently who are losing sleep because they're finding building yet another feature with "just one more prompt" irresistible. My problem is - before, I'd get ideas, start something, and it would either become immediately obvious it wouldn't be worth the time, or immediately obvious that it wouldn't turn out well / how I thought. Now, the problem is, everything starts off so incredibly well and goes smoothly... Until it doesn't.
- simonw 9mo ago
  I used to have ideas and jot them down in Apple Notes and then usually forget about them entirely. Now I have an idea and jot it down in the Claude Code tab on my iPhone... and a couple of minutes later the idea is software, and now I have another half-baked project to feel guilty about for the rest of time.
  - falloutx 9mo ago
    In couple of minutes? My claude code takes like 5 mins just to wake up and write a simple plan.
    - foretop_yardarm 9mo ago
      Possibly relevant: https://github.com/anthropics/claude-code/issues/6008 https://github.com/anthropics/claude-code/issues/6008
      - falloutx 9mo ago
        Wow an issue going on for 6 months without any input from Anthropic again, not even surprised
    - barishnamazov 9mo ago
      You are taking to simonw, surely Claude have given him free super fast unlimited token access to nightly version of Claude 5.1 Opus. (just joking, your posts are great, Simon!)
nonethewiser 9mo ago
It reduces the friction of coding tremendously. Coding was usually not the bottleneck but it still took a significant amount of time. Now we get to spend more time on the real bottlenecks. Gathering requirements from end users, deciding what should be built, etc.
- irishcoffee 9mo ago
  > It reduces the friction of coding tremendously. Coding was usually not the bottleneck but it still took a significant amount of time. I don’t think I agree. How can something be both “usually not a bottleneck” that usually “takes a significant amount of time” ? > Now we get to spend more time on the real bottlenecks. Gathering requirements from end users, deciding what should be built, etc. Sounds like you might really enjoy a PM role. Either way, LLM or not, whatever gets written up and presented will have a lot of focus on a bike shed or will make the end user realize allllll the other things they want added/changed, so the requirements change, the priorities change… So now we just don’t get to do the interesting part… engineer things. If I wanted to be a PM I’d do that.
  - switchbak 9mo ago
    Just because the magical fairy helps you write things, you still need to ensure it's engineered properly. Especially at the macro level. Some day it'll handle that, but for now it's very bound to make silly decisions that you need to be on top of, especially as those compound in a large scale system.
  - nonethewiser 9mo ago
    > How can something be both “usually not a bottleneck” that usually “takes a significant amount of time” ? I dont understand what you dont understand. Is everything that takes a significant amount of time necessarily a bottleneck? That seems implied by you but makes no logical sense. The funnel into the programming work is often more difficult/time consuming/resource intensive than the programming. Also, sometimes its not as costly but should be. And insufficient time and resources were spent up front which caused the coding portion to take a lot longer than it should. In which case the programming time may appear to be the bottleneck but it was still really the funnel leading into it. > Sounds like you might really enjoy a PM role Enjoyment isn't really a factor in terms of what work needs to be done. And designing technical features isnt really a PM responsibility.
disiplus 9mo ago
This is real, so im a freelancer, i used this small invoicing platfrom to create invoices for my customers. At "work" im working on accounting systems, and erp-s. So with AI, why would i pay monthly for invoicing when i can build it myself. After i day i had invoicing working. Like the simple thing where you get PDF out. Then i started implementing doube-entry booking. And support different tax systems. And then, but we need a sales part then crm, then warehouse. Then projects to track time and so on. And now i have a full saas that i dont need and im not going to waste time on competing in that market. Now im thinking of puting it as open source.
- SoftTalker 9mo ago
  "Invoicing for freelancers" has just about as many solutions as "to do" lists or ticket systems. Just use what you built if it works, open sourcing it is likely to get zero interest among the thousands of other options.
- fragmede 9mo ago
  what would be interesting, as i find myself building the same thing, is to know about your decision matrix of the things you chose to do and build. I've even got the contractor working on my house logging his hours using this thing. I'm considering making a stripe connect integration to do payouts with this thing!
dizhn 9mo ago
> they're finding building yet another feature with "just one more prompt" irresistible. Totally my experience too. One last little thing to make it perfect or something that I decide would be "nice to have" ends up taking so much time in total. Luckily now I can access the same agent session on my phone mobile browser too so I can keep an eye on things even in bed. (Joke but not joke :D)
keybored 9mo ago
You write a lot about AI. If this is in your free time why not just take a break? If you are ten times more productive, rest for at least twice as much. I don’t get it.
- Synthetic7346 9mo ago
  I assume that if you take a break you'll have missed a lot when you come back, at the pace things are evolving. Which is OK for some people like OP but maybe not for simonw
  - keybored 9mo ago
    I meant like watch a movie.
    - fragmede 9mo ago
      That's boring. Turns out, this is fun!
      - keybored 9mo ago
        And? You can clearly burn out while having fun. See the entire context here.
krackers 9mo ago
>now but I end the day exhausted. > another feature with "just one more prompt" irresistible The llm agents as slot machines analogy seems to be getting stronger...
- Wowfunhappy 9mo ago
  I remember when Claude Code first came out. The model wasn't as good, and there was no Claude Max plan, everything was API pricing. It really felt like a slot machine. Maybe if I try again, this feature will work. It's only another couple of dollars...

by simonw 9mo ago

bob_theslob646 9mo ago

Throw in the fact that clawdbot can work 24/7. It reminds me of why people wanted financial markets to be 24/7. We as a society should probably take a look at that otherwise it may lead to burnout in a not so small percentage of people

billylo 9mo ago
We should ask how the traders manage this. It's essentially 24/7 markets in the world. For them, the FOMO effects are even stronger... actual money earning opportunity.
nextlevelwizard 9mo ago
Why we as a society should give a fuck if someone can’t stop prompting? Unless you mean we as a society should make you pay for the damages your prompts are doing to nature?
- bob_theslob646 9mo ago
  >Why we as a society should give a fuck if someone can’t stop prompting? It is not prompting, it is the constant feeling that you always have to be "on."
  - nextlevelwizard 9mo ago
    Still dont see how that is problem rest of us should worry. Instead of you just having ounce of self control

onlyrealcuzzo 9mo ago

> I've had conversations with people recently who are losing sleep because they're finding building yet another feature with "just one more prompt" irresistible. My problem is - before, I'd get ideas, start something, and it would either become immediately obvious it wouldn't be worth the time, or immediately obvious that it wouldn't turn out well / how I thought. Now, the problem is, everything starts off so incredibly well and goes smoothly... Until it doesn't.

simonw 9mo ago
I used to have ideas and jot them down in Apple Notes and then usually forget about them entirely. Now I have an idea and jot it down in the Claude Code tab on my iPhone... and a couple of minutes later the idea is software, and now I have another half-baked project to feel guilty about for the rest of time.
- falloutx 9mo ago
  In couple of minutes? My claude code takes like 5 mins just to wake up and write a simple plan.
  - foretop_yardarm 9mo ago
    Possibly relevant: https://github.com/anthropics/claude-code/issues/6008 https://github.com/anthropics/claude-code/issues/6008
    - falloutx 9mo ago
      Wow an issue going on for 6 months without any input from Anthropic again, not even surprised
  - barishnamazov 9mo ago
    You are taking to simonw, surely Claude have given him free super fast unlimited token access to nightly version of Claude 5.1 Opus. (just joking, your posts are great, Simon!)

nonethewiser 9mo ago

It reduces the friction of coding tremendously. Coding was usually not the bottleneck but it still took a significant amount of time. Now we get to spend more time on the real bottlenecks. Gathering requirements from end users, deciding what should be built, etc.

irishcoffee 9mo ago
> It reduces the friction of coding tremendously. Coding was usually not the bottleneck but it still took a significant amount of time. I don’t think I agree. How can something be both “usually not a bottleneck” that usually “takes a significant amount of time” ? > Now we get to spend more time on the real bottlenecks. Gathering requirements from end users, deciding what should be built, etc. Sounds like you might really enjoy a PM role. Either way, LLM or not, whatever gets written up and presented will have a lot of focus on a bike shed or will make the end user realize allllll the other things they want added/changed, so the requirements change, the priorities change… So now we just don’t get to do the interesting part… engineer things. If I wanted to be a PM I’d do that.
- switchbak 9mo ago
  Just because the magical fairy helps you write things, you still need to ensure it's engineered properly. Especially at the macro level. Some day it'll handle that, but for now it's very bound to make silly decisions that you need to be on top of, especially as those compound in a large scale system.
- nonethewiser 9mo ago
  > How can something be both “usually not a bottleneck” that usually “takes a significant amount of time” ? I dont understand what you dont understand. Is everything that takes a significant amount of time necessarily a bottleneck? That seems implied by you but makes no logical sense. The funnel into the programming work is often more difficult/time consuming/resource intensive than the programming. Also, sometimes its not as costly but should be. And insufficient time and resources were spent up front which caused the coding portion to take a lot longer than it should. In which case the programming time may appear to be the bottleneck but it was still really the funnel leading into it. > Sounds like you might really enjoy a PM role Enjoyment isn't really a factor in terms of what work needs to be done. And designing technical features isnt really a PM responsibility.

disiplus 9mo ago

This is real, so im a freelancer, i used this small invoicing platfrom to create invoices for my customers. At "work" im working on accounting systems, and erp-s. So with AI, why would i pay monthly for invoicing when i can build it myself. After i day i had invoicing working. Like the simple thing where you get PDF out. Then i started implementing doube-entry booking. And support different tax systems. And then, but we need a sales part then crm, then warehouse. Then projects to track time and so on. And now i have a full saas that i dont need and im not going to waste time on competing in that market. Now im thinking of puting it as open source.

SoftTalker 9mo ago
"Invoicing for freelancers" has just about as many solutions as "to do" lists or ticket systems. Just use what you built if it works, open sourcing it is likely to get zero interest among the thousands of other options.
fragmede 9mo ago
what would be interesting, as i find myself building the same thing, is to know about your decision matrix of the things you chose to do and build. I've even got the contractor working on my house logging his hours using this thing. I'm considering making a stripe connect integration to do payouts with this thing!

dizhn 9mo ago

> they're finding building yet another feature with "just one more prompt" irresistible. Totally my experience too. One last little thing to make it perfect or something that I decide would be "nice to have" ends up taking so much time in total. Luckily now I can access the same agent session on my phone mobile browser too so I can keep an eye on things even in bed. (Joke but not joke :D)

keybored 9mo ago

You write a lot about AI. If this is in your free time why not just take a break? If you are ten times more productive, rest for at least twice as much. I don’t get it.

Synthetic7346 9mo ago
I assume that if you take a break you'll have missed a lot when you come back, at the pace things are evolving. Which is OK for some people like OP but maybe not for simonw
- keybored 9mo ago
  I meant like watch a movie.
  - fragmede 9mo ago
    That's boring. Turns out, this is fun!
    - keybored 9mo ago
      And? You can clearly burn out while having fun. See the entire context here.

krackers 9mo ago

>now but I end the day exhausted. > another feature with "just one more prompt" irresistible The llm agents as slot machines analogy seems to be getting stronger...

Wowfunhappy 9mo ago
I remember when Claude Code first came out. The model wasn't as good, and there was no Claude Max plan, everything was API pricing. It really felt like a slot machine. Maybe if I try again, this feature will work. It's only another couple of dollars...