If you're not then you will end up frustrated because it will be as if you have an over-eager trainee that is only slightly less dumb than a box of rocks when it comes to actual thinking but that has near perfect recall about a massive amount of information. A sort of idiot-savant.
I gave it a try for a practical problem to see how long it would take to converge on a solution. It did not, because its reasoning powers simply are not that strong. But it gave me an endless runaround about solutions that this time really would work and never mind all the previous times we did the same thing and it did not work. Between the over enthusiastic tone, the ridiculous confidence level and the jumping-to-conclusions mode that you just simply can't switch off I found the whole exercise more than a little bit frustrating. It never, for even a second expressed any doubt. It never warned when what it responded with was at best extremely low confidence and possibly very wrong (or even dangerous). And this was for a problem where I already knew the solution, it was in some ways entertaining to see the AI blunder about while in full possession of all of the information to see the solution.
'Great, that narrows it down'
'Now I see the problem'
'Got it'
'Ah - perfect'
etc, etc...
If your job involves actual thinking I don't believe you are in danger just yet.
Having experienced plenty of this but also having regularly experienced Claude Code writing correct, useful programs - it is reasonably possible that coding agents are the beginning of a disruption in how we code. They'll continue to improve (and we'll continue to surround them with tools and languages that feed that improvement cycle).
Combining early, faulty "worse is better" technology with a change in cost structure (tokens vs. programers) and capability (programs in minutes not days) often wins. Even when it sucks at first.
Yes, but who will write the code that trains the AI on future stuff? Someone somewhere will have to write original code, and without writing a lot of code that new original code will be of lesser quality than what went before. Once that cycle starts there will be less money to be made in programming, so fewer high grade programmers (and we already have too little of those as it is).
If you're not then you will end up frustrated because it will be as if you have an over-eager trainee that is only slightly less dumb than a box of rocks when it comes to actual thinking but that has near perfect recall about a massive amount of information. A sort of idiot-savant.
I gave it a try for a practical problem to see how long it would take to converge on a solution. It did not, because its reasoning powers simply are not that strong. But it gave me an endless runaround about solutions that this time really would work and never mind all the previous times we did the same thing and it did not work. Between the over enthusiastic tone, the ridiculous confidence level and the jumping-to-conclusions mode that you just simply can't switch off I found the whole exercise more than a little bit frustrating. It never, for even a second expressed any doubt. It never warned when what it responded with was at best extremely low confidence and possibly very wrong (or even dangerous). And this was for a problem where I already knew the solution, it was in some ways entertaining to see the AI blunder about while in full possession of all of the information to see the solution.
'Great, that narrows it down'
'Now I see the problem'
'Got it'
'Ah - perfect'
etc, etc...
If your job involves actual thinking I don't believe you are in danger just yet.