HNHacker News
TopNewBestAskShowJobs

bhagyeshsp

15 karma · joined June 15, 2022

Self-taught developer, hacker, tech-enthusiast and business owner.

What am I up to: https://bhagyeshpathak.com/now

submissionscomments
bhagyeshsp··on You can just do things
Ah yes, no doubt about it.

I hadn't thought about compounding at the time of writing this essay. One of the reasons was that I just felt like grabbing people by their necks and thrust their face in the screen while shouting "see...see..you can just do things.. don't wait.."

Sometimes, it is better to take the immediate, most feasible action: doing the thing in front of us. Then doing the next thing. Then the next..

bhagyeshsp··on Ask HN: Anyone using AI agents for active learning sprints? Here's my setup
No worries. Just created one:

https://gist.github.com/bhagyeshsp/510f66ae3d8eb3f826dfd8273...

Modify as you find fit.

bhagyeshsp··on Show HN: Flocklist, minimalist graph-based task tracker
I mostly prefer using pc to explore apps etc. Tried this a few minutes back. Got quite confused.

Just read your text here, now that makes sense. It is made for your taste, solving a particular problem. If it is working for you, great! If you want more people to adopt it, it is a different discussion altogether.

bhagyeshsp··on Biology is a Burrito: A text- and visual-based journey through a living cell
> Miles of DNA aren't getting reproduced in a few minutes at that speed.

I didn't get the "miles of DNA" reference. A single strand of DNA is approx 3 meter in length when uncoiled. Now I'm thinking how many strands may be replicated at a time.

bhagyeshsp··on Ask HN: commands.sh
Can anyone request to add `Esc` key to go back to the homepage?

But of course, for that the developer will need to make each command open in a modal instead.

bhagyeshsp··on I over-engineered my simple AI backend: distillation, router, embedding etc.
Hi HN,

I was creating an AI chat companion for one of my products and this is the compilation of my decisions and reflections. Plenty of technical parts that you'd like to look into.

*Things I think worth highlighting*

1. Cloudflare Workers

2. Custom static site for interface

3. Full system prompt at the beginning: 17000 tokens -> Ultimately 2500 tokens

4. Tried two LLMs: one as a router LLM to inject selective context + another as the main model

5. Router was a bad idea. Main issue: 6000 ms latency

6. Product matched Embedding approach perfectly (I didn't know at the time, learnt it)

7. Context distillation was a huge learning: remove all semantic words from the prompt

8. Used Promptfoo for benchmarking

9. Cosine similarity score has to be understood for your own data, don't take any random score from the internet (internet and AI suggested 0.7 similarity score, mine turned out around 0.2-0.25)

10. Found OpenAI-4o-mini the best conversation model for my case

---

Questions? Any fellow travellers having gone through the same pain?

bhagyeshsp··on Ask HN: Anyone using AI agents for active learning sprints? Here's my setup
On the same lines: "hi, let's get started" and it begins understanding my setup and as per the initialization protocol, it has the latest state:

This is the main hook in AGENT.md file:

   ## 3. Initialization Protocol (Read This First)

   Upon being activated in this directory, you MUST:

   1. **Read `CURRENT.md`** to understand the current topic under study. The current topic's content and files are under its dedicated directory. **Ignore** all other topic directories.

   2. **Read the topic-specific `README.md`**: (e.g., in `/mcp/`) to understand the specific mental models and technical constraints of the current subject.

   3. **Read `PROGRESS.md`**: To understand exactly where the last session ended and what the "Current Frontier" is.

   4. **Acknowledge State**: Start the session by briefly confirming the current learning milestone you've identified from the files.
Since the agent makes live notes inside PROGRESS.md, no matter where you left off, it is all taken care of every time you start afresh.

Mostly, the agent concludes the session once it guages that 15-20 minute sprint-worthy concepts have been covered. Try it. If you want, I can share the full notes sans-my personal info through gist.

bhagyeshsp··on Patterns for Reducing App LLM Costs (Video Series)
You're welcome.

While you research, this is the link to my article I had written a few weeks ago on my practical experience with the whole thing. I'm going to post it on HN today.

http://sisyphusconsulting.org/case-studies/2026/04/01/scalin...

bhagyeshsp··on Biology is a Burrito: A text- and visual-based journey through a living cell
Wow, aptly named "The Machinery of Life".
bhagyeshsp··on GitHub Is Sinking
The Co-pilot slop is quite irritating.

I read about the alternatives. But has anyone tried anything that resembled Github Actions? Github Pages, I can take care of, but what about Actions? They're quite helpful to me in my day to day.

bhagyeshsp··on HERMES.md in commit messages causes requests to route to extra usage billing
Who knew one of the best AI companies in the world isn't using an AI to screen for these words.

There should have been a two layer approach:

1. Run regex to screen for target word

2. If positive, run the context through a cheap AI model

bhagyeshsp··on Biology is a Burrito: A text- and visual-based journey through a living cell
What a beautiful depiction. Reminds me of high fidelity 3D animation videos I used to watch about DNA replication, cell signalling etc.

One of the most fascinating parts to me was DNA transcription. The engineering is quite precise.

Found the video I was referring to: https://www.youtube.com/watch?v=7Hk9jct2ozY

bhagyeshsp··on Patterns for Reducing App LLM Costs (Video Series)
Hi Troy,

I skimmed through the outline. Will take a look at the individual videos when I'm on PC.

But I have been through this cost-saving phase. I didn't see "prompt distillation" as one of the techniques in your outline.

The idea is to reduce your fixed prompt token size such as "system prompt" by removing semantic words completely and other methods. I saw a whopping 60% decrease in my fixed prompt token budget.

Pls note, my scale is small but the technique works nonetheless.

Edit: so, have you tried it? Or if you have tried it, how did it go?

bhagyeshsp··on Replacing labor with AI shifts costs to compute – and compute is still expensive
Unless someone discovers a breakthrough technology that decreases the cost of inference exponentially.

AI in production at scale is too costly.

bhagyeshsp··on How ChatGPT serves ads
That's an interesting idea. I think ultimately, all AI providers will serve ads with a standard protocol. Something like Universal Commerce Protocol that Google launched a few months ago: https://developers.googleblog.com/under-the-hood-universal-c...

And the browsers will protect the protocol somehow.

← PreviousPage 2 of 2