New ask Hacker News story: I feel like I did something stupid
I feel like I did something stupid
5 by anon1253 | 2 comments on Hacker News.
In December I quit my job to start on my own, with no real funding. After working on the same project for five years I felt things were out of place. It’s probably a feeling of the times, things just didn’t feel right anymore. It was a project about applying natural language processing and machine learning technology to biomedical literature. The idea was solid, still is, in my head: apply named entity linking using ontologies, and use embeddings of those ontologies to analyze, navigate and recommend literature. Basically a search engine. It just kinda never took off. These problems are hard, harder than I expected them to be. Technically challenging, sure, marketing wise: darned near impossible. I wanted to do something on my own, disrupt the academic publishing industry. And slowly an idea formed: build an academic reference manager, integrate ActivityPub for collaboration, and then add IPFS for sharing documents. As a desktop app, so as long as there were peers: it’d be very hard to turn off. If it would be a joy to use it’d be a triple whammy: great platform for the organizing references, social features to discuss and collaborate on published and gray literature, while transparently building a peer to peer network of those files and discussions. Monetize it with a subscription fee. The idea mostly came out of sheer spite. Academic publishing felt like a scam, publishers take money from publicly funded institutes, never pay the authors, never pay the reviewers, and never contributes anything back to science except for conjured-up self important metrics like “impact”. I romanticized the idea, thinking it could be done by myself. And it be done, just not by myself in the time I have before I’ve eaten through all my savings. Just some musings of someone who got sucked into the whole “just say its AI and they’ll throw money your way”-hype. That’s not how that goes. Somewhat stuck now.
5 by anon1253 | 2 comments on Hacker News.
In December I quit my job to start on my own, with no real funding. After working on the same project for five years I felt things were out of place. It’s probably a feeling of the times, things just didn’t feel right anymore. It was a project about applying natural language processing and machine learning technology to biomedical literature. The idea was solid, still is, in my head: apply named entity linking using ontologies, and use embeddings of those ontologies to analyze, navigate and recommend literature. Basically a search engine. It just kinda never took off. These problems are hard, harder than I expected them to be. Technically challenging, sure, marketing wise: darned near impossible. I wanted to do something on my own, disrupt the academic publishing industry. And slowly an idea formed: build an academic reference manager, integrate ActivityPub for collaboration, and then add IPFS for sharing documents. As a desktop app, so as long as there were peers: it’d be very hard to turn off. If it would be a joy to use it’d be a triple whammy: great platform for the organizing references, social features to discuss and collaborate on published and gray literature, while transparently building a peer to peer network of those files and discussions. Monetize it with a subscription fee. The idea mostly came out of sheer spite. Academic publishing felt like a scam, publishers take money from publicly funded institutes, never pay the authors, never pay the reviewers, and never contributes anything back to science except for conjured-up self important metrics like “impact”. I romanticized the idea, thinking it could be done by myself. And it be done, just not by myself in the time I have before I’ve eaten through all my savings. Just some musings of someone who got sucked into the whole “just say its AI and they’ll throw money your way”-hype. That’s not how that goes. Somewhat stuck now.