HNHacker News
TopNewBestAskShowJobs

kshitij_libra

26 karma · joined May 1, 2014

I build. Management, LLMs and ML are skills i’m honing.

Blog: kshitijbanerjee.com Email me at ksh.dce@gmail.com, I won’t bite.

submissionscomments
kshitij_libra··on Ask HN: What are you working on? (March 2025)
This seems very useful. Why not make it paid ? Do you think your customers won’t buy ? Have you tried ?

What would your customers need to make them want to pay for it ?

kshitij_libra··on You might not need Redis
Do you plan to write the book ? I’d like to read it
kshitij_libra··on IMG_0001
IMG_0163 - Gold
kshitij_libra··on Ask HN: Do you already see the impact of LLMs on the job prospects for dev's?
nopes.. beyond starter projects, raw usage of chatgpt hardly leads to any gains in productivity in the enterprise setting .

So far no, but in the future with more specific and enterprise suitable tooling - likely

kshitij_libra··on Ask HN: Independent Research, Review and Publication
Yes, but then a PhD is a chicken and egg problem. You can't get into the top PhD programs under a respectable professor - unless you have some publications under your name.

So how does one get published, enough to get into a good PhD program?

kshitij_libra··on Ask HN: Can LLM's produce useful code?
Right. I think the ability to run tests/compilation as a feedback mechanism is definitely interesting .

I wonder whether the problem is that its landed on an incorrect output sample (i.e re-sampling another code output would fix it ) .. vs … a feedback loop on the compiler output is what it really needs ..

kshitij_libra··on Ask HN: Can LLM's produce useful code?
No, I meant to ask HN, but didn’t realise that if I put a URL it doesn’t start a discussion.. and now there is no delete button
kshitij_libra··on Ask HN: Can LLM's produce useful code?
> Other tasks have an element of looping which is really fundamental in computer science

Have you tried any of these agent based workflows that claim to be able to loop with some amount of reasoning ?

kshitij_libra··on Ask HN: Can LLM's produce useful code?
100% I’d rather have specialised LLMs that can solve some hairy problems like CSS , than what we have now
kshitij_libra··on Ask HN: Can Devin genuinely replace the roles of developers?
Precisely. I think it helps with smaller/mundane tasks (that it has seen in its training), but the tasks that actually require a higher level reasoning and understanding of the bigger picture - are not something we can expect the current LLM's to do.

However as I was researching, there are a few interesting ideas in this space that might help these LLM's solve more complex problems in the future. Post here if interested: https://kshitij-banerjee.github.io/2024/04/30/can-llms-produ...

kshitij_libra··on Can LLM's produce useful code?
After using code LLMs' for a bit, I wonder whether there is significant productivity gain in using them. I found that its limited for me.

Main reason: a) to find the right code, I have to keep sampling it, and b) it doesn't seem to be able to solve larger / more complex problems that I actually find more need for.

I found some interesting research on combining planning-algorithms for complex problems, and some ideas on guiding the LLM's decoding process towards correctness by optimizing it via reward functions and reducing the search space. I've detailed and summarised the main points in the post above.

Questions:

1) Do you find code LLM's really useful? Please share some stories / examples where they help vs they didn't. I'm trying to form a better understanding they are just fancy, vs actually productive and useful to most

2) Any other research ideas being pursued in this field ? / what are you trying ?

kshitij_libra··on [dead]
Hi - ( some questions as i’m curious about your entrepreneurship journey )

Does anyone pay for this ? how does this compare to other products ? what made you build this ?

kshitij_libra··on More Agents Is All You Need: LLMs performance scales with the number of agents
This trend really needs to die. If you can’t come up with an original paper name , maybe the contents aren’t that original either
kshitij_libra··on Scientists trap krypton atoms to form one-dimensional gas
Is this how you kill superman?
kshitij_libra··on Ask HN: Companies that reject Agile cultism?
Meta, Shopify - both companies reject the Agile + Microservice cult
kshitij_libra··on Ask HN: How would you build OSS Figma?
yeah, I think I like your approach, but the multiplayer aspect seems to be a great add-on, as opposed to a must-have feature to get the product started.

Both Go and Rust, and wasm are new to me - so i'm wondering about whether one is superior to the other.

I think figma has a lot of feature, and penpot seems a great replica of it, but I'm not too sure if the SVG + Clojurescript is a great way to build it..

I'm really looking for the core figma feature set to build. 1. Ability to have a canvas with zooms / freedom to create elemnets 2. Components and instances, and tree structured layers 3. Autolayout + CSS options. Frames and Pages

I think with the above 3, there's a viable basic feature set that would get one off the ground.

I did try the figma plugin apis, to extend native figma capabilities - but there's a limit to things you can do with it and it's getting to a point where more investment in the figma plugin seems a bit wasteful

kshitij_libra··on Ask HN: How would you build OSS Figma?
Yeah, PenPot (mentioned in the comments below), seems to have chosen SVG as their choice of rendering. But, looking at a few customer reviews - the performance issue seems to be noticable with the SVG approach (zooms are not clean, and more than 20 screens it starts feeling slow).

Wasm seems appropriate. Why Go, instead of Rust ?

kshitij_libra··on Ask HN: How would you build OSS Figma?
nice! Are you affiliated/ working at penpot?
kshitij_libra··on Ask HN: What are you reading these days?
The Great Mental Models - Shane Parrish
kshitij_libra··on Ask HN: Any good book recommendations on distributed systems design?
https://github.com/papers-we-love/papers-we-love/tree/master...

Thank me later

kshitij_libra··on It’s Not Too Late to Quit Social Media
How to quit YouTube addiction . Social media is still doable imo but YouTube is just a dopamine doze I can’t seem to quit
kshitij_libra··on Show HN: A side project in medical devices
And motivation . The exposure to the problem gives you motivation to solve it . If someone randomly gave this idea, one could probably just ignore it ..
kshitij_libra··on Update on Our Planned Move from Azure to Google Cloud Platform
Assuming a buy, gitlab becomes part of gcp offering in the coming future ? :D
kshitij_libra··on Death of a startup founder (2016)
I understand that’s the general idea, and that’s what every one out there says , but secretly I hope for some killer advice on how not to kill yourself doing a startup . Coz life’s short . Maybe there is some strong advice that would be really helpful
kshitij_libra··on Death of a startup founder (2016)
Are there good examples of entrepreneurs who are able to maintain a healthy balance ? How does one balance life and startup (especially on initial stages)
kshitij_libra··on The end of the cloud is coming
Well, I think intrinsic ineffeciency of storing all that data inside your central and then running huge spark jobs to crunch through them creates huge costs in the first place that probably needs to be offset by higher margins being charged from the end consumer. If you can just monitor the users engagement with your offering and charge on that without having to actually go through countless logs to do that, might be what the author is leaning to. But yeah, easier said than done. Knowing which metrics to capture back centrally itself could be a challenging thing to do, if the data is decentralized ....
kshitij_libra··on The end of the cloud is coming
I think the point would be that the central server `could` gather aggregated data / important business metrics that it could use to provide further value. I think the author also mentions something like this. Ex: provide analytics back to the end user that could be a paid service, or use that information to provide a second set of b2b offerings to other partners. etc.
kshitij_libra··on What if we made video conferencing work?
Apart from bigbluebutton are there any other open-source video conferencing solutions out there ?

Rather than so many different versions of flawed product. I wonder how things would work out if a lot of creative individuals solved and made better the same problem.

Also, I think another pain point is that webcam and audio setup is tricky to most `windows` pc average user, is there a product that helps figure out the quality score of your setup and potentially help with it ?

kshitij_libra··on Mine3D: A Web 3D Minesweeper Game
Haha Awesome!

Reminds me of a 3D 2048 game I made some years back.

https://kshitij-banerjee.github.io/Cubiks-2048/

Check the dismember view I provided (space bar). That open up the cube to give a better spatial understanding of the board.

That is the problem with 3D games, I realized. Trying to visualize is a pain to most people and deviates them from the game. This was my attempt at solving that.

I think in your case, hovering over a cell and highlighting its adjacent numbers might help create better visual understanding of the state.

kshitij_libra··on How to Learn New Things as an Adult
Because we’re all getting `dumber in the age of Google`

What's wrong with that ? I fail to understand ? If you start putting everything into your brain, you'll soon go mad. A regular cleanup of the information in your brain to retain only what's relevant for your survival is okay. It's best such things are left googleable. What is the use of keeping in memory the capital of australia, when that trivia contributes a total `0` to do what you do.

If you keep doing, implementing and executing stuff that you are passionate about, I think you'll keep learning. Only thing is that we become more focussed as adults about our interests, than we are as children (inquisitive about everything).

Page 1 of 2Next →