If you want to build something instead, I suggest taking a look at indie hackers to see what other people are doing or using a tool like glimpse to find trends before they pop-off and build a solution to those things. Source: about 3 years ago

The most valuable and impressive thing you can do is build a business. Hands down. Especially if it makes money. That will show you're not just a cog in the wheel but able to critically think and have valuable practical skillsets. I would experiment with something that has tailwinds. Like an AI business, or a VR business once the new apple VR app store comes out. you'd be shocked at how much you can make... Source: about 3 years ago

For example; trends.co is not very good because the people that write for them are journalists, not business owners so although the writing is good, the ideas are poorly researched. On the other hand, meetglimpse.com is pretty good, they have nuanced and unique business ideas that you can take advantage of but the market research behind it is a little lacking, their chrome extension tool is great tho. And then... Source: over 3 years ago

a lot of things. I've built 3 online businesses that were profitable with under $1000. it's really just a hustle once you get product market fit. Starting something online shouldn't take that much money if you know how to test it. Check out like trends.vc, explodingideas.co, meetglimpse.com etc. They may be able to spark your creativity for ideas that could be good opportunities for the price point. Source: over 3 years ago

Great idea. You should scrape ideas from meetglimpse.com, explodingideas.co, trends.co and the other sites that post ideas. Would be an easy way to bulk up the document. Source: over 3 years ago

This is really cool. The annotation-to-testing-to-annotation-etc. Feedback loop makes a ton of sense, and I'd encourage others who may be confused on this post to look at the Automotus case study https://encord.com/customers/automotus-customer-story/ for the annotation side, but my understanding is the relationship between model outputs and annotation steering is out of scope for that project - do you know of... - Source: Hacker News / over 2 years ago

Spacy [0] is a state-of-art / easy-to-use NLP library from the pre-LLM era. This post is the Spacy founder's thoughts on how to integrate LLMs with the kind of problems that "traditional" NLP is used for right now. It's an advertisement for Prodigy [1], their paid tool for using LLMs to assist data labeling. That said, I think I largely agree with the premise, and it's worth reading the entire post. The steps... - Source: Hacker News / almost 3 years ago

Prodigy AI - Offers software engineers career coaching, skill assessment, and job matching. Visit Prodigy AI. - Source: dev.to / about 3 years ago

I essentially want to use a Combo of OCR + NER to attempt to identify this, but I'm not sure NER is well suited for this, as it is not natural language, so there is little context to go off of. I was thinking of perhaps using Prodigy, a data annotation tool, to annotate Candidate Names, Races, etc, and perhaps it will be able to learn off of image data alone wheat these fields tend to look like. Source: over 3 years ago

I come from a similar application area, where I try to tag (annotation/label) a taxonomy of products iteratively. You are trying something slightly different, AFAIU, labeling a flat set of songs, each song with a set of tags from ontology (directed graph)From an application point of view, this is what taxonomists often do, when migrating products from one catalog to another: mapping one taxonomy to another. There... Source: over 3 years ago