735 karma · joined June 22, 2023
At a minimum, there should be precommit checks and CI workflows that cause PRs to fail if the documentation is not up-to-date and synced with the other docs.
Then regular codebase analysis for improvement. This is where you find the bug sources, make new modules for consolidation, and get those +5000/-4000 PRs that people stuck in the world of manual code review hate.
GPT-5.6 didn't try to fix the bugs (as instructed) but it did surface them, which is something that didn't happen with GPT-5.5. When spec/plans approved Fable sends back to GPT-5.6 for ralph implementation and it seems to be an even faster, more reliable workhorse than it already was in GPT-5.5. Overall, impressed. Will continue to be a core piece of my workflow.
For example, Panini (sports card manufacturer) did Dutch auctions on boxes of new card sets during the peak of pandemic collectible mania. The majority of customers that were willing to pay the highest prices on Panini's website were card breakers, which are people/companies that sell "spots" in livestream box openings (i.e. customers buy the right to all cards containing players from a certain sports team before the box is opened).
You have to buy a smart TV and disable the internet + external device like Apple TV or Nvidia Shield.
There's no reason Aristotle or Veratad should see who the underlying requestor is. Yoti should receive the verification request, strip the context, make the request - that's it. The fact that it isn't structured that way and they are tagging on additional metadata suggests per-query economics, which creates a direct incentive to route more verifications through more parties, exactly backwards from data minimization. I'm not going to call it a rev share, but the architecture is consistent with one.
Ran gemma-4-26B-A4B-it-GGUF:Q4_K_M just fine with llama.cpp though. First time in a long time that I have been impressed by a local model. Both speed (~38t/s) and quality are very nice.
Two months later it was CVE after CVE.
All SaaS-built MCP servers should cover the entirety their existing API functionality. I know it sounds like a lot but I really don't think it is an unreasonable expectation.
A fun, brightly colored, relatively inexpensive, Windows-less laptop that you can use for doing your taxes while watching a movie has appeal. The performance isn't that important, so long as it is as responsive as the owner's phone.
Absolutely terrible products to work with both from a user and developer standpoint, but once they are up and running they are built like tanks.