Hacker News: eddyzh

New comment by eddyzh in "What happens if an entire class of workers loses faith in their careers"

eddyzh — Fri, 07 Aug 2026 20:53:28 +0000

This resonates a lot with me, well worth the long read. And the last part about how to try keep the human part relevant, is exactly what I try for now.

New comment by eddyzh in "The LLM Critics Are Right. I Use LLMs Anyway"

eddyzh — Fri, 17 Jul 2026 15:52:02 +0000

This checks out 50-250kg CO2 are common guesses for 10k spend.

That would be in the same order of magnitude of an average daily commute with an older car.

New comment by eddyzh in "Ryanair passenger sucked toward broken window after midair engine failure"

eddyzh — Sat, 11 Jul 2026 12:24:17 +0000

The fact that is is ai generated.

New comment by eddyzh in "Claude Fable is relentlessly proactive"

eddyzh — Fri, 12 Jun 2026 16:23:59 +0000

For that test you have to compare letting a fresh agent (subagent) or the same model do the same review.

The fact that a review helps does not prove the model choice for the review.

You reviewing your own writing helps too!

New comment by eddyzh in "Long-running Agents (Summary of the current state on many fronts)"

eddyzh — Wed, 29 Apr 2026 19:53:53 +0000

Added the () part since this title has been used a lot in other articles.

Long-running Agents (Summary of the current state on many fronts)

eddyzh — Wed, 29 Apr 2026 19:53:53 +0000

Article URL: https://addyosmani.com/blog/long-running-agents/

Comments URL: https://news.ycombinator.com/item?id=47953622

Points: 2

# Comments: 1

New comment by eddyzh in "Ask HN: Can you tell the difference between Claude Sonnet and Opus?"

eddyzh — Mon, 27 Apr 2026 12:18:31 +0000

Exactly this.

This may be worth the discount. Or not if your time and attention is worth (quite) a lot.

New comment by eddyzh in "Ask HN: Can you tell the difference between Claude Sonnet and Opus?"

eddyzh — Mon, 27 Apr 2026 12:09:14 +0000

At work I use opus max Fast It hardy ever fails for no reason even if I forget to give it all the right context. At home i run sonnet, and it does not get what I meant or expected 20-35% of the time. Due to the enormous difference in cost, depending on the value of your time (hourly rate) that might be a nett benefit.

Sonnet being faster alone would not be worth the failure rate for me.

At home i just not want to pay more than 20 bucks for incidental projects.

And opus max would just consume my tokens in one round.

New comment by eddyzh in "Flickr: The first and last great photo platform"

eddyzh — Sun, 26 Apr 2026 07:38:54 +0000

Thank you. This is great to hear!

New comment by eddyzh in "Graphs that explain the state of AI in 2026"

eddyzh — Sun, 19 Apr 2026 08:40:39 +0000

While chatGPT was not out then, the ML that drives robotics was acting by then very much.

New comment by eddyzh in "Chuck Norris has died"

eddyzh — Fri, 20 Mar 2026 17:34:36 +0000

In had one app like that from Cydia Loved it.

New comment by eddyzh in "Claude Code deletes developers' production setup, including database"

eddyzh — Sat, 07 Mar 2026 19:18:59 +0000

Original artikel https://open.substack.com/pub/alexeyondata/p/how-i-dropped-o...

The Mythical Agent-Month

eddyzh — Thu, 19 Feb 2026 16:30:16 +0000

Article URL: https://wesmckinney.com/blog/mythical-agent-month/

Comments URL: https://news.ycombinator.com/item?id=47075544

Points: 4

# Comments: 0

New comment by eddyzh in "Local AI is driving the biggest change in laptops in decades"

eddyzh — Wed, 24 Dec 2025 22:14:34 +0000

Would that be unified memory? Where the gpu and cpu can share the memory? Which is key for performance.

New comment by eddyzh in "Dummy's Guide to Modern LLM Sampling"

eddyzh — Sun, 04 May 2025 21:23:40 +0000

LLM are way older. The Nobel prize for it shows how they made many of the breakthroughs decades ago ChatGTP was the popular breakthrough. Even then your Smartphone keyboard has been using an LLM for a decade.

New comment by eddyzh in "Here are the Attack Plans That Trump's Advisers Shared on Signal"

eddyzh — Mon, 31 Mar 2025 05:56:32 +0000

Because optimizing for security is sometimes specifically decreasing convenience. In the original (first) article some of those considerations are discussed.

New comment by eddyzh in "Tern AI's low-cost GPS alternative actually works"

eddyzh — Mon, 17 Mar 2025 13:19:21 +0000

So I was curious what they mean by sending out an id for gps. Since that seems to me not something that is happening for all types of GPS implementations.

Tern AI's low-cost GPS alternative actually works

eddyzh — Mon, 17 Mar 2025 13:17:05 +0000

Article URL: https://techcrunch.com/2025/03/14/tern-ais-low-cost-gps-alternative-actually-works/

Comments URL: https://news.ycombinator.com/item?id=43388268

Points: 3

# Comments: 3

New comment by eddyzh in "DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL"

eddyzh — Sun, 26 Jan 2025 09:43:11 +0000

This is great context for the cost claim. Which turns out only to be technically true when looking at the final run.

New comment by eddyzh in "Most people don't care about quality"

eddyzh — Fri, 03 Jan 2025 07:46:29 +0000

>But if there was a way to correctly tell people: "look, this smartphone is 20% more expensive, but it will last twice as long and it will be more convenient for you in ways you can't understand right now", nobody would go for the worse quality, right?

I am not sure if we are living in the same world. No most people absolutely do not make rational choices like that. No matter how you tell them.

Most importantly what you think is best for people is stil a perspective. Especially in the context of tv.