With tools like Devin and Magic on the block and more constant layoffs, companies and enterprises want to see how much can they get away with reducing their workforce and go with GPT-4, GPT-5, and the like, even as an experiment.
Software engineers are still needed but not as many as back in the pre-2020, ZIRP (Zero interest rate policy) era when money was cheap and companies were burning or raising more money for a series D onwards from VCs.
I predict that with the layoffs, some software developer jobs to work with enterprises would require an 'AI certification' to work with.
This may not bode well with all startups but on top of SOC Type I, II certifications as a requirement that enterprises need from startups, across the applicant pool some startups that want to get enterprise customers may require a developer to have an 'AI certification'.
When I say this, I mean something like 'OpenAI Certified' or 'OpenAI Certified Developer'.
What's your thoughts on this?
The trouble w/ commercial AI certifications is that they’ll have a short shelf-life. Sure GPT-4 is pretty good now but as people get to understand how LLMs work you will be able to get equivalent or better performance at 10% or 1% of the inference cost and all Open AI can do is reengineer their product radically to be competitive. If it succeeds you have new stuff to learn, if it fails it is like getting a certification in using a film camera.
The real certification in A.I. is to get a MEng or PhD. The MEng will occupy you for a year and hopefully bring you right up to the present, the PhD takes so long the scene is going to be very different when you get out of school.
They are basically corporate money grabs.