@neovintage.bsky.social
Wrote up some thoughts on AI agents after digging into a research paper and comparing it with what I’ve seen firsthand at Benchling.
The pattern is the same: we keep talking about “autonomy,” but most of the work still hinges on all the human context.
neovintage.org/posts/why-ai...
The pattern is the same: we keep talking about “autonomy,” but most of the work still hinges on all the human context.
neovintage.org/posts/why-ai...
November 18, 2025 at 5:42 PM
Wrote up some thoughts on AI agents after digging into a research paper and comparing it with what I’ve seen firsthand at Benchling.
The pattern is the same: we keep talking about “autonomy,” but most of the work still hinges on all the human context.
neovintage.org/posts/why-ai...
The pattern is the same: we keep talking about “autonomy,” but most of the work still hinges on all the human context.
neovintage.org/posts/why-ai...
Fell down a rabbit hole reading that HN thread (news.ycombinator.com/item?id=459...) about “SlopStop,” opened the Kagi smallweb repo that was referenced in the comments… and, BAM!, my blog is sitting right there in smallweb.txt (github.com/kagisearch/...). Totally unexpected. Made my night.
November 14, 2025 at 5:45 PM
Fell down a rabbit hole reading that HN thread (news.ycombinator.com/item?id=459...) about “SlopStop,” opened the Kagi smallweb repo that was referenced in the comments… and, BAM!, my blog is sitting right there in smallweb.txt (github.com/kagisearch/...). Totally unexpected. Made my night.
Been on a kick reading AI papers lately to see where software development is heading. I started writing down what I've been reading from these papers. This time it was about property based testing. Having an LLM doing it over a whole codebase feels too brute force neovintage.org/posts/agenti...
October 22, 2025 at 10:39 PM
Been on a kick reading AI papers lately to see where software development is heading. I started writing down what I've been reading from these papers. This time it was about property based testing. Having an LLM doing it over a whole codebase feels too brute force neovintage.org/posts/agenti...
Frameworks do make a difference. I know you can move fast with using minimal frameworks like express or flask. But, if your app gets sufficiently large, that starting point will become a liability.
December 3, 2024 at 5:42 AM
Frameworks do make a difference. I know you can move fast with using minimal frameworks like express or flask. But, if your app gets sufficiently large, that starting point will become a liability.