• 2 Posts
  • 205 Comments
Joined 3 years ago
cake
Cake day: June 10th, 2023

help-circle











  • I can’t imagine the battery life with a second screen will be great, don’t see any mention of it on the site.

    Shame they don’t have an option for another battery. The Retroid is a great device, but better docking support for an enhanced TV experience would be preferable. Especially since then you’re already wired, and have more real estate. You see this same thing popularized from Valve and Nintendo, which I’m sure have done a large amount of research into this already.



  • I’ve been building projects with Supabase the last few years now. It’s been pleasantly surprising.

    The only feedback I have for them is:

    • the CLI tooling around migrations is lackluster
    • they don’t let you have control over the version of pgbouncer that’s running
    • the auth product offering is a sore thumb in a variety of ways (fortunately they’re beginning to fix pieces here and there such as the new getClaims method)
    • Postgres 17 support is taking a while (but looks like it’ll come out this year)



  • Building AI tools will also require very few of the skills of a manager from our generation. It’s better to be a prompt engineer, building evals and agentic AI than it is to actually manage. Management will be replaced by AI, it’s turtles all the way down. They’re going to expect you to be both a project manager and an engineer at the same time going forward, especially at less enterprising organizations with lower compliance and security bars to jump over. If you think of an organization as a tree structure, imagine if the tree was pruned, with fewer branches to the top, that’s what I imagine there end goal is.



  • If this is more than a hobby and you want it to be competitive, you’re going to have to use a lot of reliable hard data. E.g. the trends on PC part pickers site. Subjective decisions on parts will be scrutinized to a ridiculous level, so best to just stick to the numbers whenever you can, unless you have some actual data you could use to back up some deviation from the norm.

    If it needs to be accurate you may also want to get into scraping, it’s pretty easy to scrape pricing data live, just need to pay for proxies to do it at scale, or pull from an existing API if you trust them to be accurate.

    In general I think you’ll find most people already use PC part picker who have this problem, I’ve been using them for 14 years. But it sounds like you may want to automate the shopping experience they offer?