Hacker News: kovek

New comment by kovek in "LearnVector – Andrew Ng's AI company building one‑to‑one learning experiences"

kovek — Wed, 29 Jul 2026 16:49:09 +0000

Try being an English speaker learning Mandarin from Claude

New comment by kovek in "LearnVector – Andrew Ng's AI company building one‑to‑one learning experiences"

kovek — Wed, 29 Jul 2026 04:51:09 +0000

I think learning from another human being live who can understand your confusingly articulated questions and who can adjust their answers for you, another human being, that's great.

New comment by kovek in "Show HN: Remux – an open-source tmux workspace designed for iPhone"

kovek — Thu, 23 Jul 2026 17:56:24 +0000

Please support /voice for claude!

New comment by kovek in "Advertise in ChatGPT"

kovek — Wed, 22 Jul 2026 01:42:04 +0000

I think it's easy to lie to consumers. Then, how much agency do they have?

New comment by kovek in "A new era for software testing"

kovek — Fri, 12 Jun 2026 19:21:02 +0000

I have only one thing to say....

what

New comment by kovek in "A new era for software testing"

kovek — Thu, 11 Jun 2026 22:50:14 +0000

I heard people say this before. I'm wondering, how do you instruct the LLM to generate the tests? Do you tell it the scenarios that would be covered, or do you just tell it to write tests for the code?

New comment by kovek in "Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable"

kovek — Thu, 11 Jun 2026 02:46:21 +0000

I thought it was known since a few years now that if you train models to NOT do certain things, then they start behaving in weird ways…

New comment by kovek in "Headway Therapy Patients Forced to Scan Their Faces to Keep Getting Care"

kovek — Fri, 29 May 2026 18:04:23 +0000

What if the organization who tries to verify sends a request on an app on the user’s iPhone (or whatever device can do the same), and the user scans their face with FaceID to produce a file send to the organization, which will then send that file to Apple to ask if the file represents the right person? I trust Apple so that works for me.

New comment by kovek in "Taking a walk may lead to more creativity than sitting, study finds (2014)"

kovek — Tue, 26 May 2026 17:21:16 +0000

I’d say it’s possible to have creativity when you’re sitting as well. I like to think that’s it’s all about staying active. Reading, diarying, calling a friendind. All of that.

New comment by kovek in "OpenAI and Government of Malta partner to roll out ChatGPT Plus to all citizens"

kovek — Sun, 17 May 2026 02:15:57 +0000

LLMs are like a search engine that autocompletes. It's a tool.

New comment by kovek in "Meta to receive $3.3B in tax breaks for its $10B Louisiana data center"

kovek — Fri, 15 May 2026 21:41:33 +0000

What negative consequences does being unelected have?

New comment by kovek in "Natural Language Autoencoders: Turning Claude's Thoughts into Text"

kovek — Fri, 08 May 2026 04:21:15 +0000

Maybe you can't 100% know what every layer "thinks", if you go through all the layers, you might see a cohesive "thinking" story. So, if there is any information you lose at layer N, you might learn some of it in layer N+1. The masking in the layers is not deterministic so the model can't really consistently lie throughout the layers. It doesn't chose what information we get to inspect. There might be a game of whack-a-mole, but you might get a general sentiment. I think the more layers there are, the more the model itself can hide very nuanced lies (But by that time we'd have a better mind-reading model).

However, I haven't read about it yet. I'm really excited to look into it!

New comment by kovek in "The Self-Cancelling Subscription"

kovek — Thu, 07 May 2026 20:42:49 +0000

I’ve read recently about natural systems in the book Antifragile. It’s interesting how those systems can become better.

New comment by kovek in "SubQ: a sub-quadratic LLM with 12M-token context"

kovek — Wed, 06 May 2026 04:59:48 +0000

> The core idea is content-dependent selection. For each query, the model selects which parts of the sequence are worth attending to, and computes attention exactly over those positions.

I don't know if this will help for things like understanding code, where the all relevant parts can be the file of 1000 lines that we are analyzing, and where every token is relevant in understanding recursion, loops, function calls, etc.

This sounds like it would be great to do SSA before passing things along to a code model like claude code.

Let me know if I misunderstood

New comment by kovek in "We decreased our LLM costs with Opus"

kovek — Wed, 29 Apr 2026 17:42:00 +0000

I don’t think triaging is necessarily an easy task

New comment by kovek in "We decreased our LLM costs with Opus"

kovek — Wed, 29 Apr 2026 06:44:16 +0000

Does thinking about how to offload matter?

New comment by kovek in "An update on recent Claude Code quality reports"

kovek — Fri, 24 Apr 2026 18:21:34 +0000

10s of GBs? ( 1,000,000 context * 1,000 vector size ) ^ 2 = 1,000,000,000,000,000,000… oh wow.. I must be miscalculating

What about only storing the conversation and then recomputing the embeddings in the cache? Does that cost a lot? Doing a lot of matrix multiplication does not cost dollars of compute, especially on specialized hardware, right?

New comment by kovek in "An update on recent Claude Code quality reports"

kovek — Thu, 23 Apr 2026 21:05:30 +0000

What if the cache was backed up to cold storage? Instead of having to recompute everything.

New comment by kovek in "Show HN: How I Topped the HuggingFace Open LLM Leaderboard on Two Gaming GPUs"

kovek — Tue, 10 Mar 2026 16:51:51 +0000

Is this similar to send 48656c6c6f2c20686f772061726520796f753f in the prompt? As done here: https://youtu.be/GiaNp0u_swU?si=m7-LZ7EYxJCw0k1-

New comment by kovek in "Ask HN: How to be alone?"

kovek — Sun, 08 Mar 2026 21:27:45 +0000

> You seem to be hinting at the "chemical imbalance" theory of antidepressants, which has been largely debunked

Can you say more?