HN user

vasili111

1,475 karma
Posts107
Comments515
View on HN
news.ycombinator.com 2y ago

Ask HN: What you are thinking about?

vasili111
4pts5
news.ycombinator.com 2y ago

Ask HN: What are you thinking about?

vasili111
1pts1
news.ycombinator.com 2y ago

Ask HN: Cloud solutions architects specialized on ML and data engineering?

vasili111
2pts0
news.ycombinator.com 2y ago

Ask HN: How much math you need to work as MLE in a company but not academia?

vasili111
1pts0
news.ycombinator.com 2y ago

Ask HN: Which AI/DL/ML/DS related discord channels do you recommend to join?

vasili111
3pts1
news.ycombinator.com 3y ago

Ask HN: Which AI tools do you use?

vasili111
11pts12
news.ycombinator.com 3y ago

Ask HN: Which AI tools you are using for programming and how it helps you?

vasili111
1pts1
news.ycombinator.com 3y ago

Ask HN: One curved monitor vs. 2 or 3 non-curved for software development

vasili111
7pts17
www.youtube.com 4y ago

Why Did MIT Switch from Scheme to Python? (Gerald Jay Sussman)

vasili111
3pts0
www.nature.com 5y ago

Quantum circuits with many photons on a programmable nanophotonic chip

vasili111
1pts0
reviverestore.org 5y ago

The world’s first successfully cloned endangered Przewalski’s horse

vasili111
2pts0
science.sciencemag.org 5y ago

Universal coherence protection in a solid-state spin qubit

vasili111
1pts0
phys.org 5y ago

Scientists discover way to make quantum states last 10k times longer

vasili111
3pts0
news.ycombinator.com 6y ago

Hacker News Lists

vasili111
2pts0
dosbox-staging.github.io 6y ago

Dosbox-Staging 0.75.0 Release

vasili111
1pts0
cran.r-project.org 6y ago

R 4.0.0 Released

vasili111
1pts1
stackoverflow.blog 6y ago

My Most Embarrassing Mistakes as a Programmer (So Far)

vasili111
1pts0
en.wikipedia.org 6y ago

Inverted Pyramid (Journalism)

vasili111
1pts0
www.youtube.com 7y ago

We Should Invest in Rat Massage

vasili111
1pts0
www.eurekalert.org 7y ago

Deep-learning approach predicts protein structure from amino acid sequence

vasili111
2pts0
www.spynumbers.com 7y ago

Spy numbers stations that may be on air now

vasili111
3pts0
edition.cnn.com 7y ago

China is exporting the Great Firewall

vasili111
2pts0
www.sciencemag.org 7y ago

A famous mathematician today claimed he has solved the Riemann hypothesis

vasili111
1pts0
www.plymouth.ac.uk 7y ago

Scientists receive $1.3M to study new propulsion idea for spacecraft

vasili111
1pts1
www.amusingplanet.com 7y ago

Green Bank: The Town with No Cell Phones, Wi-Fi or Radio

vasili111
2pts0
security-tracker.debian.org 7y ago

Linux Kernel TCP Implementation Vulnerable to Denial of Service

vasili111
3pts1
www.youtube.com 7y ago

Introduction to Freenet

vasili111
2pts0
www.youtube.com 8y ago

How I Manage Resources in Pure C

vasili111
7pts1
www.healththoroughfare.com 8y ago

Higgs Boson Observed Breaking Down into B Quarks

vasili111
1pts0
greatlakesledger.com 8y ago

A two bottom quarks decay has been spotted within the Higgs boson

vasili111
2pts0
How I Use "AI" 2 years ago

If I know technology which I am using llm for then llm helps me to do it faster. If I am not familiar with technology then llm helps me to learn it faster by showing me win the code that it generates which part of technology is important and how it works in real examples. But I do not think it is helpful and I would say it may be dangerous depending on task you do if you do not know technology and also do not what to learn it and understand how generated code works.

AI is not a bubble and already bring real benefit and will bring even more in future. It will spread more and at some point it will became so common that we will just use it and just talk less about it. For example, same happened with mobile phones, at first everybody was talking about them but now everybody just use them and does not talk much about it.

Links that you have provided are good for learning how deep learning in general and LLMs in particular work. But if you are interested in only building the products based on existing models (like GPT models from OpenAI) you will not need those details of inner work and how those model are created. In that case learn from OpenAI documentation, Azure OpenAI documentation, Azure AI services documentation and etc.

Large Language models are pre-trained by creators on the huge data.

In many cases you do not need to do anything with LLM and you can just use it.

If they were not trained on the data that contains information that you are interested then you can use technique called RAG (Retrieval-Augmented Generation).

You also can do fine-tuning which is kind of training but on small amount of data.