Hacker News: tsunego

New comment by tsunego in "Geography of generative AI workforce impact to differ from previous technologies"

tsunego — Thu, 06 Mar 2025 18:31:49 +0000

AI is expensive so no surprise

New comment by tsunego in "QwQ-32B: Embracing the Power of Reinforcement Learning"

tsunego — Thu, 06 Mar 2025 18:30:05 +0000

Can’t wait to see if my memory can even acocomodate this context

New comment by tsunego in "Richard Sutton and Andrew Barto Win 2024 Turing Award"

tsunego — Wed, 05 Mar 2025 19:45:34 +0000

> This has nothing to do with his professional life.

you mean his personal life?

New comment by tsunego in "The Differences Between Deep Research, Deep Research, and Deep Research"

tsunego — Wed, 05 Mar 2025 14:20:14 +0000

Love me some good old whataboutism (sure, LLMs are now super-intelligent at writing software, but can they clean my kitchen? No? Ha!)

New comment by tsunego in "The Differences Between Deep Research, Deep Research, and Deep Research"

tsunego — Wed, 05 Mar 2025 14:18:22 +0000

Neat summary but you forgot Grok!

New comment by tsunego in "AI: Where in the Loop Should Humans Go?"

tsunego — Tue, 04 Mar 2025 22:16:21 +0000

Great reminder that AI is basically overpriced copilot: impressive enough to sell, flawed enough to need babysitting, and guaranteed to make sure humans still get blamed when things inevitably go wrong.

GRPO Judge Experiments: Findings and Empirical Observations

tsunego — Tue, 04 Mar 2025 22:07:23 +0000

Article URL: https://kalomaze.bearblog.dev/grpo-judge-experiments-findings-and-empirical-observations/

Comments URL: https://news.ycombinator.com/item?id=43260309

Points: 2

# Comments: 0

New comment by tsunego in "Enhancing Frame Detection with Retrieval Augmented Generation"

tsunego — Tue, 04 Mar 2025 14:02:38 +0000

Cool paper but unsurprising results since anything benefits from RAG

New comment by tsunego in "Show HN: Agents.json – OpenAPI Specification for LLMs"

tsunego — Mon, 03 Mar 2025 17:55:14 +0000

I like your approach but it's not clear to me whether it's MCP compatible

Anthropic just announced a MCP registry

New comment by tsunego in "An Attempt to Catch Up with JIT Compilers"

tsunego — Mon, 03 Mar 2025 16:31:33 +0000

chasing inline cache micro-optimizations with dynamic binary modification is a dead end. modern CPUs are laughing at our outdated compiler tricks. maybe it's time to accept that clever hacks won’t outrun silicon.

New comment by tsunego in "GPT-4.5: "Not a frontier model"?"

tsunego — Sun, 02 Mar 2025 15:57:02 +0000

Exactly! designed specifically for people who love burning corporate budgets.

New comment by tsunego in "Show HN: I built a modern Goodreads alternative"

tsunego — Sun, 02 Mar 2025 15:54:59 +0000

No offense, but you're completely missing why Goodreads is still relevant.

People aren't sticking around for shiny features or slick UI—they stay because Goodreads has a critical mass of users and reviews.

The value isn't in half-stars or fancy shelves; it's in the network effects. Unless you have a way to bring over millions of active reviewers (and their reviews), you're just building another pretty ghost town.

New comment by tsunego in "GPT-4.5: "Not a frontier model"?"

tsunego — Sun, 02 Mar 2025 15:50:34 +0000

GPT-4.5 feels like OpenAI's way of discovering just how much we'll pay for diminishing returns.

The leap from GPT-4o to 4.5 isn't a leap—it's an expensive tiptoe toward incremental improvements, priced like a luxury item without the luxury payoff.

With pricing at 15x GPT-4o, they're practically daring us not to use it. Given this, I wouldn't be surprised if GPT-4.5 quietly disappears from the API once OpenAI finishes squeezing insights (and cash) out of this experiment.

New comment by tsunego in "Mark Cuban offers to fund former 18F employees"

tsunego — Sun, 02 Mar 2025 15:46:50 +0000

Privatization isn't about eliminating human flaws but accountability and incentives. Private corporations face real consequences if they're inefficient: customers leave, profits drop, and they eventually fail. Government agencies, funded by taxpayers without alternatives, rarely face such direct pressures.

New comment by tsunego in "The independent researcher (2018)"

tsunego — Tue, 25 Feb 2025 15:55:34 +0000

It's funny how this was written in 2018 but thanks to AI in 2025 this is becoming increasingly accessible.

In fields like AI research, tools such as Cursor and Claude/GPT are effectively mini-research assistants that help refine hypotheses and accelerate coding experiments.

While fields requiring expensive experimental resources (psychology, medicine, physics) remain challenging, even these areas benefit from AI-powered analysis of observational data.

The barriers to meaningful contribution are lower than ever.

New comment by tsunego in "Launch HN: SubImage (YC W25) – See your infra from an attacker's perspective"

tsunego — Mon, 24 Feb 2025 21:24:59 +0000

Actionability >>> observability

If you can pull this off, you will have a great time

New comment by tsunego in "MKWS – Simple Static Site Generator"

tsunego — Mon, 24 Feb 2025 17:52:37 +0000

website takes a long time to load for me (which is ironic for a static website generator)

my 2 cents are that the problem with static websites IMHO is not to generate some HTML but to have some inspirational templates

hugo has a ton of cool templates one can fork quickly and adjust

New comment by tsunego in "Show HN: Bracket City – A daily, exploded (?) crossword puzzle"

tsunego — Mon, 24 Feb 2025 17:51:01 +0000

cool idea, especially the newsletter so i don't need to visit the website every day

New comment by tsunego in "Introduction to Stochastic Calculus"

tsunego — Mon, 24 Feb 2025 17:48:31 +0000

still wild to me that diffusion models are fast becoming the secret sauce behind ai image generation, but their roots are buried deep in stochastic calculus

who knew brownian motion would eventually help create cat memes?

New comment by tsunego in "Microsoft cancels leases for AI data centers, analyst says"

tsunego — Mon, 24 Feb 2025 16:03:25 +0000

No stargate?