Hey there,
Happy (belated) International Sloth Day! š¦„
Week 3 of the Hacktoberfest Open-Source AI Challenge is officially live, and the theme is Slow and Steady. Your mission: build something with open-weight models or open-source AI that rewards laziness. Think a local agent that summarizes your group chats overnight, an offline bedtime story generator, or a tool that finally tidies the inbox you keep putting off. Let the code do the work so you donāt have to.
Thereās $2,850 in prizes up for grabs across 21 winners. Submissions are due October 25 at 11:59 PM PDT.
Looking ahead, our final Hacktoberfest challenge, kicks off Monday, October 26 with a brand-new theme. Register now so youāre first to know when details drop.
Lastly, make sure to tune into the livestreams we will be hosting this week, right on the DEV homepage:
Thursday, October 22
- 10:30AM - 11:15AM ET: Making Open Models Smarter with Tools, MCP, and Skills.
- 1:00PM - 1:45PM ET: What Happens When You Call a Model API?
- 2:00PM - 3:00PM ET: Open Source Ready Live: Hacktoberfest Edition
Friday, October 23
- 11:00AM - 11:45AM ET: Build an Open-Source Software Factory Powered by Open Models.
Have a great week ā¤ļø
As always, here are our favorite posts from last week in no particular order:
- AI torture chamber: deep dive into whether model can feel pain: one developerās hands-on exploration of the research behind a viral repo and what āpainā really means inside a model.
- Natural Languages Have Design Trade-offs Too: Georgian Is Dynamically Typed, English Is Statically Typed: a native Georgian speakerās take on how human languages mirror the design trade-offs of programming languages.
- What I Actually Learn When I Build With AI: personal insights into the gap between shipping working code with AI and truly understanding it.
- The Bug Lives at the Seam, Not in the Logic: lessons learned from a field test where every unit test passed and the bugs showed up anyway.
- What 9,050 AI-Written Scripts Taught Me About Slop: real experiences from benchmarking 181 model configurations on matching one writerās voice.
- Eight Minutes to Live. Was It Cursor, or Was It the Spec?: a minute-by-minute look at building and shipping a dual-currency tip calculator with an AI agent.
- Can a local LLM replace Claude Code? The benchmark everyone quotes is the wrong one.: rethinking which benchmarks matter when running local models as coding agents.
Here are some standout wins from last week's discussion:
- @dj29: took their last internal college exam and solo-hosted their first Hacktoberfest Meetup š¤
- @smbmunna: landed a Junior Frontend Engineer role after six months of trying, pivoting from six years in implementation and support š¼
- @maame-codes: shipped a seriously good creamy prawn pasta for dinner š
- @ale3oula: wrapped up onboarding and finished their first real week at a new job š
- @chrisvasqm: made a comeback to DEV after more than four years away š
Well done, everyone!
Thanks for Being Awesome! š
Have we told you lately how awesome you are? Thanks a ton for all you bring to DEV. Your ideas, your content, your supportāit's what makes this community great.
Happy coding!