I just created a pretty complex script in ChatGPT ...
# suitescript
c
I just created a pretty complex script in ChatGPT which was easily 90% correct.
🫠 3
a
I'll paste what I said in a previous thread... as a tech consultant if the value you're bringing to the table is that you generate code, then you're doing it wrong. Writing the code isn't the hard part, its finding the correct solution, understanding the requirements, and even recommending requirement changes to provide 80% of what a client needs but only taking 20% of the time that the 100% solution would.
c
@Anthony OConnor You're alluding to the difference between a transactional developer and a consultant. The difference is highly dependent on the organisational structure. Some will sit at the end of a pipeline of designs and requirements of which they have no input, others will be actual consultants.
a
I've never had a role as dev where I was handed a requirements doc and technical spec that was set in stone. Its true that different orgs and different roles have different levels of freedom afforded to developers, but regardless of that as you advance in your career you should be moving closer to the decision making process no matter where you start.
c
I've seen SuiteScript devs with many years of experience that sit at the end of the pipeline I described, it tends to be the case at the bigger (multi-billion dollar) customers only though.
a
that's very surprising to me, I've worked at billion dollar companies, and had clients that were, and never saw this. lots of redtape... some things that were set in stone with no way to provide input/change hearts and minds... but even in those big places that wasn't the norm
c
It's standard in any large waterfall bureaucracy. Mileage may vary of course.
It's exactly how I started my Java and SAP career at Accenture, the dev in the corner is a code monkey and that's all. Only the architects and BAs got to drive requirements and solutioning.
s
If ChatGPT could generate code that uses libraries effectively such as NFT and lodash, I'd be interested. Until then, it can keep its amateur code.
s
that is interesting @michoel - not quite correct but better than I'd expected
I hadn't even heard of phind until now
I just tried that exact same question again on phind and it returns garbage compared to the example you linked. Bah.
m
Make sure you use best model option (GPT-4)
It's amazing, it's same engine as ChatGPT (GPT-3.5 / 4) but optimised for programming questions