Ask HN: Are you worried your non-AI project might soon be obsolete? • Hacker News Board

Re: Ask HN: Are you worried your non-AI project might soon be obsolete?

#21

Post by richardjam73 » Sat, Apr 15, 2023, 1:11 PM UTC

mdmglr wrote: post #18

richardjam73 wrote: post #13

Sometimes I feel like I live in an entirely different universe from other HN users and posts like this reinforce that view.

Can you elaborate more?

It is mostly about peoples motivations for programming and the complexity of their projects. Although I would not mind making money off of my projects that is not why I start them, also my projects are often quite complicated and don't revolve around web tech that much.

Re: Ask HN: Are you worried your non-AI project might soon be obsolete?

#22

Post by SeanAnderson » Sat, Apr 15, 2023, 3:24 PM UTC

Yeah, kinda. It's not even really that related, but I've been tinkering around with indie game development and I just can't shake the feeling that the quality of experience waiting to be delivered through emergent interactions with AI agents is going to make whatever I cobble together look trite in comparison. Time will tell, though.

Re: Ask HN: Are you worried your non-AI project might soon be obsolete?

#23

Post by james-revisoai » Sun, Apr 16, 2023, 3:42 PM UTC

vmc_7645 wrote: post #14

I'm even worried about my AI project because these larger companies just have so much more in terms of resources. Unless I find some novel methodology that's difficult to replicate without its explicit capability my product is dead in the water.

Same boat. I do fear this. Like you it sounds, I have worked on an app that was data-intensive and used older machine learning techniques(well from 2018-2021... BERT, T5 models, OCR, since outdated), and main existing competitors being able to deploy GPT cheaply undoubtedly negatively hit our business.

It also decreases in my opinion, the likelihood and value of acquisitions of startups, at least those easily substituted with an LLM.

I think there is strategy or tangents where LLMs aren't as effective, but these are harder to demo or convince investors of. it's a tricky time.