These 3 AI models tried to build me a database

Поделиться
HTML-код
  • Опубликовано: 6 янв 2025

Комментарии • 10

  • @llucis-v
    @llucis-v Месяц назад +3

    Excellent presentation, Greg! Big fan of your work -- btw cool shout-out to the "rabbit hole" (syndrome? 😉) about the blog post at the end there 👏-- and indeed the blog post for this is an absolute treasure, the details there are amazing, such a great read and rich learning experience!

  • @good-dev-student
    @good-dev-student Месяц назад +2

    Supabase team awesome 🎉❤

  • @Techonsapevole
    @Techonsapevole 29 дней назад

    weird, isn't qwen2.5-coder:7b better for sql and tool calling ?

  • @SALAVEY13
    @SALAVEY13 Месяц назад

    cool stuff! i was recreating my whole schema with it, it is awesome - but at some point migration file stopped being updated and responses started to appear in chat as code snippets, still good, made a ru video on my channel right away about it - i haven't yet reached token limit, amazing!

  • @riccardoesclapon549
    @riccardoesclapon549 Месяц назад +1

    This is amazing! For me it doesn't seem to work well however. Using openrouter exactly as shown I get "There was an error connecting to your custom model provider: {"success":false,"error":"Unauthorized"}" but I tested the API key and it works. I also don't see any of the tables be generated when it responds. I also tried deploying it but not surprisingly there are no tables in the new deployment

    • @thomaseruchie1262
      @thomaseruchie1262 Месяц назад

      wrere u able to resolve it

    • @SALAVEY13
      @SALAVEY13 Месяц назад

      @@thomaseruchie1262 workaround would be collect all sql snippets in one giant file and feed it into command line in supabase, like recreate migration file manually, haven't tried yet, hope it'll work - i ran out free projects, gonna try local installation and i would have to use file way either way)

  • @e_v_a_l_ca
    @e_v_a_l_ca Месяц назад +4

    Been using Claude from Windsurf and it’s been hit and miss. Great timing on this video. Personally, I would be hesitant to use any AI’s out of China

    • @Techonsapevole
      @Techonsapevole 29 дней назад

      Local LLM doesn't communicate with servers outside your computer

  • @CraigCannon
    @CraigCannon Месяц назад +1

    🔥