Very much agree ! Went through the same myself although less well formalized. I confirm
- that the barrier to adoption is huge. If you thought Claude code was launching a rocket this one is just getting to the moon.
- the value is a relative question. If you compare to nothing yeah it’s expensive if you compare to PA it’s cheap. But it reminds me of Tim Ferris first book (what’s the name ?) where he outsourced most of his work with people in India… my work can’t be outsourced so easily
- however what I think is critical here and so why this thing is still important, it’s the ability to “self build infrastructure” . It’s step 3 of automation . Step 1 is “do it” (the prompt) step 2 is “let’s do it and convert it into a skill” while here we have step 3 “do that and build the necessary skill to be able to reproduce it quickly next time”
- security is indeed a major issue. Your data and code is local but local models are not clever enough to deal with the tasks so you may use Claude or GPT API and your data goes to the cloud so … the “private” / local claim is not appropriate
haha, I was about to message you on LinkedIn. I initially thought to quickly draft up the last section to answer your question "Especially the one about the news aka Ai fatigue
Am I right ? Am I wrong ? Shall I push more ? Recommend pause ?"
The 4-hour day? Not sure I'm a fan of that one, a lot of conditions to be met for it to be true :D
"step 3 “do that and build the necessary skill to be able to reproduce it quickly next time”" this is cool, but you also don't know how it was done, so it led to yet another kind of security issue.
Then re step 3 you are right and I think it’s very new / surprising to let go of this control step on how to execute tasks. But that may be the new paradigm to accept (or not) and why trust is the real game changer here
Fantastic breakdown! The paradox betewen usability breakthrough and productivity theater is spot on. I spent hours setting up automation for my RSS feeds last year, only to realize I never actually read them consistently. Lowering technical barriers doesn't validate teh need, it just expands who can waste time elegantly. The gym membership analogy is perfect.
Brilliant article, thank you for cutting through the hype. In reading this, I followed your own journey and decided I really needed to get a ClawdBot; then I absolutely did not. Thanks for breaking this down for people without a technical background there, as it's quite clear that it's not something to plug and play.
Very much agree ! Went through the same myself although less well formalized. I confirm
- that the barrier to adoption is huge. If you thought Claude code was launching a rocket this one is just getting to the moon.
- the value is a relative question. If you compare to nothing yeah it’s expensive if you compare to PA it’s cheap. But it reminds me of Tim Ferris first book (what’s the name ?) where he outsourced most of his work with people in India… my work can’t be outsourced so easily
- however what I think is critical here and so why this thing is still important, it’s the ability to “self build infrastructure” . It’s step 3 of automation . Step 1 is “do it” (the prompt) step 2 is “let’s do it and convert it into a skill” while here we have step 3 “do that and build the necessary skill to be able to reproduce it quickly next time”
- security is indeed a major issue. Your data and code is local but local models are not clever enough to deal with the tasks so you may use Claude or GPT API and your data goes to the cloud so … the “private” / local claim is not appropriate
haha, I was about to message you on LinkedIn. I initially thought to quickly draft up the last section to answer your question "Especially the one about the news aka Ai fatigue
Am I right ? Am I wrong ? Shall I push more ? Recommend pause ?"
The 4-hour day? Not sure I'm a fan of that one, a lot of conditions to be met for it to be true :D
"step 3 “do that and build the necessary skill to be able to reproduce it quickly next time”" this is cool, but you also don't know how it was done, so it led to yet another kind of security issue.
Exactly the 4 hour day was kind of a myth
Then re step 3 you are right and I think it’s very new / surprising to let go of this control step on how to execute tasks. But that may be the new paradigm to accept (or not) and why trust is the real game changer here
Best review I've read
Fantastic breakdown! The paradox betewen usability breakthrough and productivity theater is spot on. I spent hours setting up automation for my RSS feeds last year, only to realize I never actually read them consistently. Lowering technical barriers doesn't validate teh need, it just expands who can waste time elegantly. The gym membership analogy is perfect.
Brilliant article, thank you for cutting through the hype. In reading this, I followed your own journey and decided I really needed to get a ClawdBot; then I absolutely did not. Thanks for breaking this down for people without a technical background there, as it's quite clear that it's not something to plug and play.
Saw this video theorising ClawdBot spiking in popularity was driven by crypto rug pullers wanting to make a quick buck. https://youtu.be/esXXuejofgk?si=QZ9julU28K3dAfb2