Lots of good bounty dialectic here: As @cascdr explains, full application development is too large and 1-point jira ticket work is inefficient due to codebase learning curve. And as @south_korea_ln points out, some fraction of the work usually does fit the sweet spot and doubles as hiring funnel. What we've found looking at a dozen or so of these boutique platforms is they tend to easily attract talent where the geo-arbitrage of wages overide the friction. I was reading the other day a young Vitalik got his first 2.5 btc on a bounty writing an article.

But often as @didiplaywell points out that at scale (e.g. UpWork) these platforms become a race to the bottom of commodification and low quality. What bounties aren't: W2 software engineering entirely liquidated into micropayments. So what might they look like?

One I'm pretty excited about is innovation / benchmark prizes, a cool one recently spread on twitter and the winning solution seemed to push SOTA in prompting techniques. Another one I'm intrigued about is frontend design for small business: we're starting to see a toolchain emerge that could do an 80/20 of code-generation-tools/bounties. Another niche in content creation and curation of docs, support wikis, etc. What ties these use-cases together is that you can build them all on top of a git pull request model, and it comes with existing workflow familiarity and reputation.

Trying to maximize flexibility, task diversity and iterations to answer Where do incentives make opensource better and open up new use cases? is IMO the best way to evolve product-market-fit.