Hacker News: mfalcon

New comment by mfalcon in "Who is Satoshi Nakamoto? My quest to unmask Bitcoin's creator"

mfalcon — Thu, 09 Apr 2026 16:46:31 +0000

I don't know what requires more skills: creating Bitcoin or avoiding getting identified after all these years.

New comment by mfalcon in "Tell HN: I'm 60 years old. Claude Code has re-ignited a passion"

mfalcon — Sat, 07 Mar 2026 01:31:34 +0000

"Just when I thought I was out, they pull me back in"

New comment by mfalcon in "Ask HN: How do you debug multi-step AI workflows when the output is wrong?"

mfalcon — Fri, 20 Feb 2026 12:07:14 +0000

You have to evaluate the llm responses. https://aunhumano.com/index.php/2025/09/03/on-evaluating-age...

Verifying coding AIs for LLM powered software

mfalcon — Tue, 03 Feb 2026 20:18:09 +0000

Article URL: https://aunhumano.com/index.php/2026/02/03/verifying-coding-ais/

Comments URL: https://news.ycombinator.com/item?id=46876670

Points: 2

# Comments: 0

New comment by mfalcon in "Ask HN: Are you going to meetups/conferences?"

mfalcon — Thu, 22 Jan 2026 12:07:47 +0000

In my opinion, conferences are mostly high noise low signal. My boss went to an "AI conference" last month and all the learnings he shared were things I was advocating already:

- use claude code

- evals are critical

- don't use AI everywhere

I think the conferences are more "work vacations" than work. Maybe they're useful if you think about them this way.

New comment by mfalcon in "Ask HN: How to bullet proof yourself from AI?"

mfalcon — Mon, 19 Jan 2026 12:25:00 +0000

The truth is that no one really knows.

My approach is to "ride the wave", don't resist, just try to use it as an advantage and if not possible, adapt and look where I can keep adding value to the world.

Currently I'm trying to get my own company started because that way I can iterate the product faster using AI and if it goes well I can slowly migrate from coder to manager. I've come to terms that my days as a pure coder are coming to an end but I've been coding for at least 15 years.

New comment by mfalcon in "Opus 4.5 is not the normal AI agent experience that I have had thus far"

mfalcon — Wed, 07 Jan 2026 10:56:11 +0000

Wake me up at Opus 12

New comment by mfalcon in "Turning an old Amazon Kindle into a eInk development platform (2021)"

mfalcon — Tue, 30 Dec 2025 15:27:08 +0000

Would be nice to use as an interface to interact with Claude Code.

New comment by mfalcon in "Project Vend: Phase Two"

mfalcon — Sun, 28 Dec 2025 01:58:31 +0000

Nice, I'll take a look. I was thinking about building a benchmark similar to the one you described, but first focusing on the negotiation between the store and the product suppliers.

Does your software also handle this type of task?

New comment by mfalcon in "Speech and Language Processing (3rd ed. draft)"

mfalcon — Mon, 15 Dec 2025 16:38:25 +0000

I was eagerly waiting for a chapter on semantic similarity as I was using Universal Sentence Encoder for paraphrase detection, then LLMs showed up before that chapter :).

New comment by mfalcon in "The "confident idiot" problem: Why AI needs hard rules, not vibe checks"

mfalcon — Mon, 08 Dec 2025 15:29:22 +0000

I had been working on NLP, NLU mostly, some years before LLMs. I've tried the universal sentence encoder alongside many ML "techniques" in order to understand user intentions and extract entities from text.

The first time I tried chatgpt that was the thing that surprised me most, the way it understood my queries.

I think that the spotlight is on the "generative" side of this technology and we're not giving the query understanding the deserved credit. I'm also not sure we're fully taking advantage of this funcionality.

New comment by mfalcon in "Evaluating Agents"

mfalcon — Thu, 04 Sep 2025 03:47:06 +0000

You can evaluate with your programming language of choice.

New comment by mfalcon in "Evaluating Agents"

mfalcon — Thu, 04 Sep 2025 03:44:17 +0000

Good idea for a follow up post :)

New comment by mfalcon in "Evaluating Agents"

mfalcon — Thu, 04 Sep 2025 03:43:54 +0000

Yes, and these problems are more present in the first iterations, when you are still trying to get a good enough agent behaviour.

I'm still thinking about good ways to mitigate this issue, will share.

New comment by mfalcon in "Evaluating Agents"

mfalcon — Wed, 03 Sep 2025 23:35:09 +0000

Hey fellow hners, OP here. Been working on agents for a while so I started sharing some things.

The idea is to keep updating this post with a few more approaches I'd been using.

Evaluating Agents

mfalcon — Wed, 03 Sep 2025 23:32:57 +0000

Article URL: https://aunhumano.com/index.php/2025/09/03/on-evaluating-agents/

Comments URL: https://news.ycombinator.com/item?id=45121547

Points: 42

# Comments: 9

New comment by mfalcon in "An LLM is a lossy encyclopedia"

mfalcon — Tue, 02 Sep 2025 14:58:08 +0000

I know that the "understanding" is a stretch, but I refer to the Understanding of the NLU that wasn't really understanding either.

New comment by mfalcon in "An LLM is a lossy encyclopedia"

mfalcon — Tue, 02 Sep 2025 12:34:23 +0000

I think that the natural language understanding capability of current LLMs is undervalued.

To understand what the user meant before LLM's we had to train several NLP+ML models in order to get something going but in my experience we'll never get close to what LLM's do now.

I remember the first time I tried ChatGPT and I was surprised by how well it understood every input.

New comment by mfalcon in "SpaCy: Industrial-Strength Natural Language Processing (NLP) in Python"

mfalcon — Wed, 27 Aug 2025 10:53:26 +0000

I guess it depends on how you use the LLMs. We implemented some workflows where the LLMs were used only for dialogue understanding, then the system response was generated by classic backend code.

New comment by mfalcon in "Getting good results from Claude Code"

mfalcon — Fri, 08 Aug 2025 17:27:26 +0000

That's the way I'd used it, I've built a document with all the requirements and then gave it to CC. But it was not a final document, I had to go back and make some changes after experimenting with the code CC built.