News for hackers. Stuff that matters.

akshay326

akshay326 on Hacker News

Show HN: Using RasPi and Qwen 2.5 VL to find out when birds visit my balcony

Posted by akshay326 on Saturday April 04, 2026 @04:22PM
from the show-hn dept.

https://akshay326.com/bird-feeder/

I had one question: when do birds show up at my bird feeder?Setup: Raspberry Pi Zero 2W + USB webcam + Supabase (free tier) + 93 lines of Python that captures a 1280x720 frame every 60 seconds with retry logic for the Pi's flaky WiFi.Analysis: Ran 1,500 images through Qwen 2.5 VL 7B via OpenRouter API. 996 tokens per image. Total cost: $0.37.Results:- 433 birds detected across 3 days (23.7% of...

( Read More... | no comments | 4 points | on HN )
Re:Open Source in the Age of AI on Saturday February 28, 2026 @01:26AM (#47191191)
Cloudflare cloned Next.js because the source was right there. Most closed products have 80% of their complexity locked behind auth, encrypted state, and prerequisite flows — you can't reverse-engineer what you can't observe
Re:Optimize_anything: A Universal API for Optimizing Any Text Parameter on Sunday February 22, 2026 @09:46PM (#47117438)
thanks for shipping gepa! read about it today morning, impressed by its prompt optimization results! PS - the github.io link is truncated
Re:Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases? on Wednesday January 28, 2026 @01:50AM (#46791890)
> if you're worried the model is reinforcing your biases..... i agree, i don't understand many domains well enough, yet i feel there's value in calling out assumptions, irrespective of how hard verification is
Re:Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases? on Tuesday January 27, 2026 @01:38AM (#46776277)
> Sometimes I go a step further and ask the question with the opposite bias..... curious to try this have you ever found it biasing you in the opposite direction tho?
Re:Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases? on Monday January 26, 2026 @07:57PM (#46774040)
wow, i wonder how bulletined & concise the outputs of your prompt might be!

have you ever felt this prompt being restrictive in some sense? or found a raw LLM call without this preamble better?

Re:Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases? on Monday January 26, 2026 @07:51PM (#46773988)
> Have the model play the Devil's advocate. i've tried this sometimes. only issue being dumb me skipping to add similar phrasing every time i open claude or gemini

have you found a way to consistently auto-nudging the model by default?

Re:Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases? on Monday January 26, 2026 @07:44PM (#46773936)
Ah interesting. i like actor-critic models! do you use it just for coding or non-technical chats too?

Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases?

Posted by akshay326 on Monday January 26, 2026 @03:15PM
from the ask-hn dept.

Lately i've been experimenting with this template in Claude's default prompt ``` When I ask a question, give me at least two plausible but contrasting perspectives, even if one seems dominant. Make me aware of assumptions behind each. ```I find it annoying coz A) it compromises brevity B) sometimes the plausible answers are so good, it forces me to thinkWhat have you tried so far?

( Read More... | 23 comments | 5 points | on HN )
Re:The 26,000-Year Astronomical Monument Hidden in Plain Sight on Tuesday January 20, 2026 @01:58PM (#46696243)
> There is an angle for doubt, for sorrow, for hate, for joy, for contemplation, and for devotion.

I’m so intrigued - what was going on inside Hansen's brain?

Re:Ask HN: How can we solve the loneliness epidemic? on Thursday January 15, 2026 @09:14PM (#46642222)
Dogs?
Re:Ask HN: Could you share your personal website here? on Thursday January 15, 2026 @03:02PM (#46638340)
way outdated but i mumble a few things every now and then -- https://akshay326.com/
Re:Ask HN: What is the best way to provide continuous context to models? on Thursday January 15, 2026 @12:38AM (#46628487)
What would you cover not-continuous?

Best methods I’ve observed -progressive loading (claude skills) & symbolic search (serena mcp)

Re:Show HN: Open-source welcome email automation: Supabase + Gmail + Seer on Wednesday January 14, 2026 @01:18PM (#46619988)
trouble self-hosting? checkout our 1-click Railway setup: https://github.com/seer-engg/seer
Re:Best way to find chill job where I can learn and grow as a swe on Tuesday January 13, 2026 @03:35PM (#46607517)
100% agreed - i think its hard to have both comfort and growth/discomfort at once. for the 'chill' part, you can sandbag and make some money at FMAANG or whatever the new acronym is
Re:Show HN: I used Claude Code to discover connections between 100 books on Monday January 12, 2026 @01:04AM (#46584663)
This is pretty cool, I wanted to do something with my Readwise Reader too. Jinx today claude code created a 3D Neo4J visualizer tool for YC advice + my quotes collection. Code here - https://github.com/akshay326/quote-viz
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Sunday January 11, 2026 @10:34AM (#46576601)
I’ve used mini for synthetic dataset generation extensively. Never tried Mistral; will check it out
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @08:43PM (#46571947)
true both - i've observed i end up spending more tokens + time with linting, than without
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @01:25PM (#46568453)
which simple models have you found good?
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @12:14AM (#46562994)
thanks for the idea! https://x.com/akshay326_/status/2009856179854561476
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @12:11AM (#46562978)
LLMs try to cheat. all sorts of evasive ways or smart tricks in some cases to avoid working on context-heavy tasks. i've constantly observed if left unchecked it tries to loosen the lint settings
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @12:08AM (#46562967)
thanks i've not used PostToolUse but will checkout. i'm excited about Rust's autofixable issues promise. curious how effective they are, and how deep of a issue can they solve
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @12:06AM (#46562953)
currently starting to do the same over seer's frontend, i didn't realise how simple yet effective this technique / guardrail could be!
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @12:04AM (#46562945)
amen! that's my bitter lesson for the time being, unless claude gets eerily better
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Saturday January 10, 2026 @12:03AM (#46562936)
i agree, the tool is indeed broken. its simultaneously stupid and smart in different ways. but i think there's some value in continuing to use and evaluate it
Re:Pre-Commit Lint Checks: Vibe Coding's Kryptonite on Friday January 09, 2026 @09:29PM (#46562162)
Accurate TL;DR. Probably should've led with that instead of burying it 380 lines deep in an autopsy report :)

Older »