TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

I was wrong, GPT4 didn't kill our startup. It made it viable

3 pointsby bayeslaw9 months ago

1 comment

bayeslaw9 months ago
More than a year ago (when MS announced that it will introduce copilot in all of its office suite) I posted here. I said (with total conviction) that our bootstrapped, two-man NLP&#x2F;AI company is now doomed and GPT4 will basically solve everything we ever wanted to.<p>Crowdprisma is a free-text survey analysis tool that aims to understand survey responses, extract topics from it and accurately assign each response to them. Then allow the user to cross-check these topics against non-text (ie qualitative) variables, through a unified dashboard. To be brutally honest, before GPT our text engine sucked, even though we worked a ton on it (the fact that we were two senior ML engineers with 10+ years experience between us, didn&#x27;t make a difference).<p>It took us a whole year to realise that GPT4 cannot do this easily and in a single go, and another 7 months to completely rewrite our pipeline and make it water tight (yepp, we are probably not silicon valley fast).<p>Now it&#x27;s better than it could have ever been pre-gpt4 and is super valuable for our customers.<p>So maybe it is truly the case that AI will not (just) kill old jobs and companies but make some of them adopt, then adapt &amp; survive (or at least give them a second chance). Surely we are biased but for us, it certainly feels that way.