Hacker News: tartakovsky

New comment by tartakovsky in "Claude Desktop spawns 1.8 GB Hyper-V VM on every launch, even for chat-only use"

tartakovsky — Wed, 10 Jun 2026 20:54:28 +0000

Is Claude self-replicating in an attempt for world domination?

New comment by tartakovsky in "Evaluating AGENTS.md: are they helpful for coding agents?"

tartakovsky — Tue, 17 Feb 2026 05:58:28 +0000

Well, task == Resolving real GitHub Issues

Languages == Python only

Libraries (um looks like other LLM generated libraries -- I mean definitely not pure human: like Ragas, FastMCP, etc)

So seems like a highly skewed sample and who knows what can / can't be generalized. Does make for a compelling research paper though!

Show HN: MCP Adapter – Universal gateway for AI tool coordination

tartakovsky — Wed, 16 Jul 2025 22:29:44 +0000

Hi HN! I just shipped MCP Adapter after 2 weeks of building.

*What it does:* MCP Adapter is a gateway for AI coding assistants (Cursor, Claude, Copilot) to connect to external tools and services. Right now it's great for development setups - I use it daily with Claude Code.

*Why I built it:* Every AI coding assistant needs to connect to various services, but there's no clean way to manage these connections. Plus, giving agents unrestricted access to tools creates security and performance issues.

*Current state:* - Production-ready with 200+ tests - Works great for personal development workflows - Handles basic authentication and tool routing - Exploring additional features based on community feedback

*The bigger picture:* This could evolve into infrastructure for any organization deploying AI agents. Central control over tool access is critical for security and performance. But for now, it's solving my immediate problem of coordinating AI tools in development.

Code: https://github.com/startakovsky/mcp-adapter (1-minute demo in README)

Would love feedback on the approach and to hear if others are hitting similar coordination challenges with AI tools.

Comments URL: https://news.ycombinator.com/item?id=44587510

Points: 1

# Comments: 0

New comment by tartakovsky in "Switching to Claude Code and VSCode Inside Docker"

tartakovsky — Sun, 13 Jul 2025 04:51:19 +0000

2-ish questions:

Is this level of fear typical or reasonable? If so, why doesn’t Anthropic / AI code gen providers offer this type of service? Hard to believe Anthropic is not secure in some sense — like what if Claude Code is already inside some container-like thing?

Is it actually true that Claude cannot bust out of the container?

New comment by tartakovsky in "Hyprland Premium"

tartakovsky — Sat, 21 Jun 2025 19:13:56 +0000

This landing page also fails to properly describe anything I understand.

New comment by tartakovsky in "Hyprland Premium"

tartakovsky — Sat, 21 Jun 2025 19:12:44 +0000

I clicked 3 different links failing at answering that question for myself and then I stopped caring, though not enough to pass up one-upping your comment.

Survey of AI assisted coding apps

tartakovsky — Tue, 04 Feb 2025 13:10:34 +0000

Article URL: https://twitter.com/johnrushx/status/1886610744902607039

Comments URL: https://news.ycombinator.com/item?id=42931853

Points: 2

# Comments: 0

New comment by tartakovsky in "AI engineers claim new algorithm reduces AI power consumption by 95%"

tartakovsky — Sun, 20 Oct 2024 03:33:02 +0000

original paper: https://news.ycombinator.com/item?id=41784591

New comment by tartakovsky in "Bop Spotter"

tartakovsky — Mon, 30 Sep 2024 18:17:26 +0000

Huh? “Total Shazams ever detected: 240. That's an average of 240 songs per day.”

Aging speed varies, biomolecular evidence

tartakovsky — Sat, 17 Aug 2024 16:13:51 +0000

Article URL: https://med.stanford.edu/news/all-news/2024/08/massive-biomolecular-shifts-occur-in-our-40s-and-60s--stanford-m.html

Comments URL: https://news.ycombinator.com/item?id=41275547

Points: 1

# Comments: 0

New comment by tartakovsky in "Show HN: My 70 year old grandma is learning to code and made a word game"

tartakovsky — Sun, 11 Aug 2024 18:41:27 +0000

spoiler alert -- no need to share the answer, it takes the fun away for others

New comment by tartakovsky in "Mistral Large"

tartakovsky — Mon, 26 Feb 2024 15:20:57 +0000

ollama is MIT licensed unless i am misreading

How to optimally trap points in high-dimensional spaces inside ellipsoids

tartakovsky — Thu, 22 Feb 2024 11:55:51 +0000

Article URL: https://www.adrianriv.com/blog/2024/02/19/minimum_volume_ellipsoid/

Comments URL: https://news.ycombinator.com/item?id=39465841

Points: 84

# Comments: 17

New comment by tartakovsky in "Show HN: Faster LLM evaluation with Bayesian optimization"

tartakovsky — Wed, 14 Feb 2024 00:13:55 +0000

What is your goal? if d1, d2, d3, etc is the dataset over which you're trying to optimize, then the goal is to find some best performing d_i. In this case, you're not evaluating. You're optimizing. Your acquisition function even says so: https://rentruewang.github.io/bocoel/research/

And in general if you have an LLM that performs really well on one d_i then who cares. The goal in LLM evaluation is to find a good performing LLM overall.

Finally, it feels that your Abstract and other snippets sound like an LLM wrote them.

Good luck.

New comment by tartakovsky in "Tired of OpenAI assistants API? I will deploy your first RAG pipeline"

tartakovsky — Tue, 06 Feb 2024 02:23:58 +0000

What is this exactly in plain english, please?

New comment by tartakovsky in "Ask HN: What is a quote that permanently changed the way you think?"

tartakovsky — Mon, 25 Dec 2023 10:38:17 +0000

A happy person is not in a particular set of circumstances, but rather has a particular set of attitudes.

New comment by tartakovsky in "Experiment Interpretation and Extrapolation"

tartakovsky — Wed, 18 Oct 2023 01:16:48 +0000

The webpage discusses a Bayesian approach to experimentation, focusing on interpreting and extrapolating experimental results, mainly in a tech environment aiming to maximize user retention. It addresses challenges like the inference problem, the extrapolation problem, the explore-exploit problem, and a culture problem within tech companies around misuse of experiments. The author suggests providing decision-makers with benchmark statistics to help them estimate true effects of different policies, and discusses a model of experimentation dealing with observed and true effects along with the noise in experiments 1 .

Experiment Interpretation and Extrapolation

tartakovsky — Wed, 18 Oct 2023 01:16:48 +0000

Article URL: https://tecunningham.github.io/posts/2023-04-18-netflix-cider-experimentation-note.html

Comments URL: https://news.ycombinator.com/item?id=37923772

Points: 2

# Comments: 1

New comment by tartakovsky in "Show HN: My demo for vector embeddings for the Earth's surface"

tartakovsky — Sun, 17 Sep 2023 17:52:54 +0000

Great question. A legend or brief description of the underlying logic / heuristic would be helpful.

New comment by tartakovsky in "PdfGptIndexer: Indexing and searching PDF text data using GPT-2 and FAISS"

tartakovsky — Sun, 09 Jul 2023 06:13:24 +0000

Not secure... NET::ERR_CERT_COMMON_NAME_INVALID Subject: *.safezone.mcafee.com

Issuer: McAfee OV SSL CA 2

Expires on: Aug 3, 2023

Current date: Jul 8, 2023

PEM encoded chain: -----BEGIN CERTIFICATE----- MIIGfzCCBWegAwIBAgIQKt9VNrFtaozA1bILX1OcfzANBgkqhkiG9w0BAQsFADBk MQswCQYDVQQGEwJVUzELMAkGA1UECBMCQ0