Hacker News: sarthakjain

New comment by sarthakjain in "I want to wash my car. The car wash is 50 meters away. Should I walk or drive?"

sarthakjain — Mon, 16 Feb 2026 14:58:07 +0000

These days LLMs refuse to back track and use their intelligence to justify their answers.

You said: I want to wash my car. The car wash is 50 meters away. Should I walk or drive?

ChatGPT said: Walk.

At ~50 meters, driving saves essentially no time once you factor in unlocking, starting, pulling out, and parking, and it adds a small but real chance of minor dings or awkward maneuvering in a tight car-wash entry. Walking is simpler and lower-friction.

You said: okay how should the car reach there?

ChatGPT said: Drive it there when you’re ready to start the wash.

Walk over first to check the queue/payment situation and confirm they’re open. If it’s clear, walk back, get the car, and drive the 50 meters into the bay. If they handle pickup, hand them the keys and walk.

Benchmarking LLMs on Document Automation

sarthakjain — Thu, 13 Mar 2025 06:29:33 +0000

Article URL: https://nanonets.com/automation-benchmark

Comments URL: https://news.ycombinator.com/item?id=43350816

Points: 2

# Comments: 0

Accel backs startup using AI to Kill Finance Paperwork

sarthakjain — Wed, 13 Mar 2024 00:36:00 +0000

Article URL: https://www.bloomberg.com/news/articles/2024-03-12/accel-backs-startup-seeking-to-use-ai-to-kill-finance-paperwork

Comments URL: https://news.ycombinator.com/item?id=39686653

Points: 1

# Comments: 0

AI vs. AE: lead qualification using LLMs

sarthakjain — Thu, 07 Dec 2023 07:48:41 +0000

Article URL: https://nanonets.com/blog/decoding-bant-ais-precision-vs-human-intuition-in-sales/

Comments URL: https://news.ycombinator.com/item?id=38553936

Points: 6

# Comments: 0

SaaS moats are changing from system of record to workflow orchestration

sarthakjain — Fri, 22 Sep 2023 07:24:57 +0000

Article URL: https://www.sarthakjain.com/p/the-shifting-software-moat-from-systems

Comments URL: https://news.ycombinator.com/item?id=37608973

Points: 3

# Comments: 0

New comment by sarthakjain in "Claude Shannon Demonstrates Machine Learning (1952)"

sarthakjain — Wed, 27 Jan 2021 01:18:30 +0000

Great observation. The solution to the update problem is relatively simple. It doesn't do a search again on update. Instead everytime it encounters an update in what it knows, it just changes the data stored in memory. All it is doing is updating its learned representation. After this it still knows what the other obstacles are without having to do DFS or BFS again. If the solution was a graph, it just deleted a edge it still knows what all the other edges are. If it encounters another change it updates the state of the graph again.

With regards to Neural Networks, if they are given a reward function, which can be dynamically evaluated (in this case did I reach the end or not) they are pretty good at learning without feedback.

New comment by sarthakjain in "Claude Shannon Demonstrates Machine Learning (1952)"

sarthakjain — Tue, 26 Jan 2021 20:25:18 +0000

It's interesting to note why this was considered AI in 1952 and some may not consider it to be AI today. The AI was the search algorithm to find an effecient solution to the maze, not the mouse being able to navigate it later in a second run. The second run was just a demonstration of it having found the solution demonstrating it's intelligence. The actual intelligence was it's first run through the maze. Almost any configuration of the maze could be solved using algorithms like depth first, breadth first and a star search (didn't check which one the video demonstrates). Even though the algorithm was trivial it's ability to be applied to problems of today is still extraordinary. Nerural networks being equally trivial algorithms capable of remarkable things. Id argue this is as much AI today as it was back then, just more people know how Shannon performed this magic trick.

New comment by sarthakjain in "Applying Apple's Framework for Avoiding Mediocrity"

sarthakjain — Mon, 11 Jan 2021 07:48:18 +0000

OP here have modified the title to remove the confusion and added an explicit note to avoid any misrepresentation.

Applying Apple's Framework for Avoiding Mediocrity

sarthakjain — Sun, 10 Jan 2021 16:58:42 +0000

Article URL: https://www.sarthakjain.com/p/apples-framework-for-escaping-mediocrity?r=86tsj

Comments URL: https://news.ycombinator.com/item?id=25716652

Points: 20

# Comments: 13

McKinsey Report - When Will the Covid Pandemic End

sarthakjain — Mon, 28 Sep 2020 10:19:07 +0000

Article URL: https://www.mckinsey.com/industries/healthcare-systems-and-services/our-insights/when-will-the-covid-19-pandemic-end

Comments URL: https://news.ycombinator.com/item?id=24615053

Points: 1

# Comments: 0

New comment by sarthakjain in "Why is AI so useless for business?"

sarthakjain — Tue, 26 May 2020 15:26:54 +0000

Not trying to just put in a baseless plug, but most of what you say can be refuted if you try out our product. Go here: https://nanonets.com/ocr-api/

Pandemic can be controlled, says Noble Laureate

sarthakjain — Tue, 24 Mar 2020 13:39:38 +0000

Article URL: https://www.thetimes.co.uk/article/pandemic-can-be-controlled-says-scientist-michael-levitt-bp7qd39lv

Comments URL: https://news.ycombinator.com/item?id=22674696

Points: 2

# Comments: 1

New comment by sarthakjain in "Table Detection and Extraction Using Deep Learning"

sarthakjain — Wed, 22 Jan 2020 10:29:09 +0000

Sudoku is a relatively simpler problem since the structure is known apriori and becomes as simple as pattern matching.

New comment by sarthakjain in "DeepSORT: Deep Learning to Track Custom Objects in a Video"

sarthakjain — Sat, 26 Oct 2019 17:54:50 +0000

Doesn't work for occlusions and needs a very high sampling rate for fast moving objects. Doesn't work at all for dense object tracking.

New comment by sarthakjain in "Building a license plate reader from scratch with deep learning"

sarthakjain — Thu, 12 Sep 2019 20:45:37 +0000

The primary idea behind this post is to show how you can use OCR technology to one of the simplest problems (LPR). Given how widespread Licence Plates are and how easily you can collect this data. Most other use cases are a little more niche and difficult to relate with.

I doubt LPR will be the primary utility of this kind of tech.

OCR in 2019 – From Character Recognition to Information Extraction

sarthakjain — Thu, 08 Aug 2019 13:35:44 +0000

Article URL: https://nanonets.com/blog/ocr-apis-to-extract-text-from-images/

Comments URL: https://news.ycombinator.com/item?id=20644387

Points: 2

# Comments: 0

Using Deep Learning to Track Objects in a Video

sarthakjain — Tue, 23 Jul 2019 14:08:00 +0000

Article URL: https://nanonets.com/blog/object-tracking-deepsort/

Comments URL: https://news.ycombinator.com/item?id=20506819

Points: 4

# Comments: 0

Using StyleGANs to Generate GoT Characters

sarthakjain — Mon, 20 May 2019 07:27:47 +0000

Article URL: https://blog.nanonets.com/stylegan-got/

Comments URL: https://news.ycombinator.com/item?id=19958203

Points: 2

# Comments: 0

Mass Production of iPhones to Start in India

sarthakjain — Mon, 15 Apr 2019 23:16:31 +0000

Article URL: https://www.bloomberg.com/news/articles/2019-04-15/foxconn-is-poised-to-begin-mass-production-of-iphones-in-india

Comments URL: https://news.ycombinator.com/item?id=19669716

Points: 245

# Comments: 99

New comment by sarthakjain in "YC Interviews in India"

sarthakjain — Thu, 28 Mar 2019 17:12:02 +0000

Typically you have ~2 weeks between being notified and interview. Have personally witnessed a lot of teams where one founder couldn't make the interview in the time frame. For the actual YC program both the invitation letter is much stronger, you probably have an established US entity and a strong reason to be in the US. Less likely you get denied a Visa.