HNHacker News
TopNewBestAskShowJobs

nnurmanov

180 karma · joined November 21, 2017

submissionscomments
nnurmanov··on Tell HN: Anthropic no longer allowing Claude Code subscriptions to use OpenClaw
That’s how ecosystems start: by cutting out third parties.
nnurmanov··on Ask HN: Founders of estonian e-businesses – is it worth it?
What about EU Inc? Is there any timeline?
nnurmanov··on Reports of Telnet's death have been greatly exaggerated
Do you have to restart your computer to exit telnet?:)
nnurmanov··on Lotus Health nabs $35M for AI doctor that sees patients for free
They claim: Since AI models are also prone to hallucinations, the company always has board-certified human doctors from top health institutions such as Stanford, Harvard, and UCSF review the final diagnoses, lab orders, and medical prescriptions., but is it economically viable?
nnurmanov··on IBM CEO says there is 'no way' spending on AI data centers will pay off
I agree. Here is my thinking. What if LLM providers will make short answers the default (for example, up to 200 tokens, unless the user explicitly enables “verbose mode”). Add prompt caching and route simple queries to smaller models. Result: a 70%+ reduction in energy consumption without loss of quality. Current cost: 3–5 Wh per request. At ChatGPT scale, this is $50–100 million per year in electricity (at U.S. rates).

In short mode: 0.3–0.5 Wh per request. That is $5–10 million per year — savings of up to 90%, or 10–15 TWh globally with mass adoption. This is equivalent to the power supply of an entire country — without the risk of blackouts.

This is not rocket science — just a toggle in the interface and I believe, minor changes in the system prompt. It increases margins, reduces emissions, and frees up network resources for real innovation.

And what if EU/California enforces such mode? This will greatly impact DC economy.

nnurmanov··on Ask HN: Quality of recent gens of Dell/Lenovo laptops worse than 10 years ago?
I used to be a Dell customer; all my family members had Dell Latitude laptops. But I agree, they got worse. I had a fan break, a key fall out, and Bluetooth issues with some of them, and when it was time to upgrade, I moved to Apple MacBook. It took some time to learn how to work in the macOS environment, but I am happy now.
nnurmanov··on Launch HN: Hypercubic (YC F25) – AI for COBOL and Mainframes
AI for COBOL? Now I’ve seen everything.
nnurmanov··on N8n raises $180M
It’s per core licensing with core factors. Core factor ranges from 0.25 to 1. I have not checked how it is now, but it used to be as above 5-10 years ago
nnurmanov··on Ask HN: How do you handle application spam via Who's Hiring?
Nowadays, when I post a job for our openings, I get thousands of applicants, which is much higher than 2–3 years ago. Over 90% are irrelevant. To limit the applicant count, I ask for references. For example, just the other day a candidate reached out directly, I asked if he could provide references, and then heard nothing back. It’s the quickest way to filter out irrelevant candidates. Now I am employing the same tactics and building an ATS where references make you stand out.
nnurmanov··on Ask HN: What's your experience with using graph databases for agentic use-cases?
I am using Neo4j to build an equipment database, I also use MySQL in the same project to store transactional data. It took some time to figure out right syntax for Spring Boot/Neo4j Cypher query, but now it works OK. The reason I chose Neo4j? Because I wanted to play with it:). I can say it is more flexible than relational databases. I would like to continue using it, but you can't create multiple instances within a database, I guess it is possible to do so by installing separate binaries, but I have not tried it yet.
nnurmanov··on Everything that's wrong with Google Search in one image
It is a business model problem, not Google’s problem. When it is easy to stand out by paying money, it will always become trash
nnurmanov··on Top Programming Languages 2025
English should be #1:)
nnurmanov··on Six months into tariffs, businesses have no idea how to price anything
I live in such country, local currency volatile, laws are changing. There is no long term planning, projects should span 1-2 years and if they bring income, then you are lucky. In long term the situation is no good
nnurmanov··on Vibechart
In the marketing world 1>2:)
nnurmanov··on DaisyUI: Tailwind CSS Components
Are there any cool UI libraries for Svelte or Vue? Similar to Baseweb or Grommet?I am planning to move away from ReactJS some day
nnurmanov··on Show HN: We moved from AWS to Hetzner, saved 90%, kept ISO 27001 with Ansible
I moved from managed AWS to unmanaged AWS (lightsail), decreasing the cost significantly and still staying in AWS ecosystem. I use S3, Route53, SES and other cheap services, you could consider this path
nnurmanov··on ZjsComponent: A Pragmatic Approach to Reusable UI Fragments for Web Development
With java I can create new package let’s say under com.examle.plugins and extend my app, namely I can add nee classes, new features and have them working with my existing app. Are there such thing with frontend?
nnurmanov··on Nanonets-OCR-s – OCR model that transforms documents into structured markdown
Are there benchmarks for such kind of tools? How does it handle tables? Different languages?
nnurmanov··on Ask HN: Any PDF Benchmarks?
For example, markitdown from MS can't recognize Cyrillic text, when I started researching into it, I found that they use pdfminer.six under the hood and there is an unresolved issue with supporting languages.

Docling is OK with tables, but fails with cyrillic text;

marker-pdf is OK with tables, but it also fails with cyrillic text;

What other pdf parser libraries exist? I am looking for preferably on-premise solutions, but if I won't find a reliable and accurate solution, I might consider cloud based solutions as well.

nnurmanov··on YAML: The Norway Problem (2022)
Another solution is to change the country name:)
nnurmanov··on CERN scientists find evidence of quantum entanglement in sheep
When the universe is born, at t=0, supposedly all the particles were entangled. Sheeps are no exception, although they are macroscopic.
nnurmanov··on Mistral OCR
I did not try doctly, but AWS Textract does not support in my case Russian, so the output is completely useless
nnurmanov··on New tools for building agents
Does anyone know if there are any difference if you typed the question with typos vs you did it correctly?
nnurmanov··on Ask HN: What Are You Building?
Let me start. I am building VoicePoll.io a voice-first survey platform where respondents record their thoughts instead of typing, making feedback effortless and natural.

Did you know that:

90% of human communication is non-verbal—intonation and pauses reveal engagement, sentiment, and authenticity.

Speaking is 3-5x faster than typing, leading to richer, more detailed responses.

Our AI-powered system transcribes, summarizes, and analyzes feedback, capturing not just words but emotions and deeper insights that text-based surveys miss.

With VoicePoll.io, you get actionable insights instantly—no manual review needed. Whether for employee feedback, customer insights, or research, our platform helps you understand what truly matters. Start recording your answers today!

nnurmanov··on Ingesting PDFs and why Gemini 2.0 changes everything
E.g. oftentimes there is l and I (capital I), this may be an issue for OCR. The perfect case is when there is a PDF document and data embedded as XML data, but unfortunately it is not the case.
nnurmanov··on Ingesting PDFs and why Gemini 2.0 changes everything
It is not OCR to blame, when you have garbage in you should not expect anything of high quality, especially with handwriting and tables and different languages. Even human beings fail to understand some documents (see doctor's prescriptions)
nnurmanov··on Ingesting PDFs and why Gemini 2.0 changes everything
prompt and pray, this is my default mode while working with LLMs
nnurmanov··on Ingesting PDFs and why Gemini 2.0 changes everything
I have not tested Azure Document Intelligence, Google Document AI, but AWS Textract, LLamaparse, Unstructured and Omni made to my shortlist. I have not tested Docling, as I could not install it on my Windows laptop.
nnurmanov··on Ingesting PDFs and why Gemini 2.0 changes everything
If you see above, someone is using a second and even third LLM to correct LLM outputs, I think it is the way to minimize hallucinations.
nnurmanov··on Ingesting PDFs and why Gemini 2.0 changes everything
This is my problem as well; do you have lots of documents?
← PreviousPage 2 of 9Next →