3 ms·
I'm more of the belief that: - Layoffs would have happened anyway, natural cycle of shedding underperformers. - the current crop of graduates have too little
by athrowaway3z 1y ago
I'm more of the belief that:
- Layoffs would have happened anyway, natural cycle of shedding underperformers.
- the current crop of graduates have too little AI experience from their education, and too little practical experience to be useful in an industry re-arranging itself.
- the average developer quality (i.e. boot camps & curriculum) has dropped the value of "a developer" from a manager's perspective, up to the point that an average Actual Indian is competitive (even when accounting for the communication overhead/loss).
Combined, that shows fewer people on the books, which they'll obviously frame in the best light; i.e. frugal & good management + innovative AI.
- whywhywhywhy 1y ago> the current crop of graduates have too little AI experience from their education Why would you need education to understand how to use the thing that means you don't need education
- rimunroe 1y agoI see fairly few people claiming effectively using AI doesn’t require learning anything. I see a lot of people claiming it requires learning a lot less than traditional methods.
- Esophagus4 1y ago> the average developer quality (i.e. boot camps & curriculum) has dropped the value of "a developer" from a manager's perspective If that’s true, I would say it’s because there are simply more developers (therefore more below average developers), not because the good developers are getting less good. If anything, my experience is that the top developers are far better now. For an internship position I’m hiring for, I’m getting resumes with several previous internships, a top 20 university’s CS degree, research lab experience, side projects, (brief) exposure to a myriad of languages and tools, and certifications. For an internship, that’s far above what I or my peers had when we were interns.
- dgfitz 1y ago> the current crop of graduates have too little AI experience from their education They probably have more "AI" (read: LLM) experience than working professionals. Not like chatGPT came out last week, they've had it their entire undergrad. Also, that's like saying "they don't have enough experience using google" which is just absurd. You type shit into a text box and see what gets shat back out. If it doesn't work, you adapt it. Prompt engineering is not hard. At best, it takes days-to-weeks learning how to prompt. I think people who are proud of themselves for spending hours carefully crafting prompts are trying to justify the time spent. It isn't hard.