yewenjie

Why isn't Android support there?

> Our first formalization of Fermat’s Last Theorem, performed by an unreleased research version of Claude, took 11 days. It relied on a purpose-built harness along with guidance from a human researcher. By contrast, a team of 1,000 Fable 5.2 agents fully formalized the theorem in just 30 hours, with minimal human intervention. Fable 5.2 started from minimal prompts, built its own ways to coordinate, and used no mathematical references, only a list of intermediate theorem statements to score progress. It needed one-third of the action tokens as our first formalization and produced a proof seven times shorter. As can be seen in the figure below, Opus 5.5 (with the same 1,000-agent setup) completed only 26% of the formalization in 18 hours, and barely made any more progress after 30.

Most people don't really have intuitions on what it means to push the frontier now (because AI has surpassed everything they are good at, for example, coding). We are truly inside the singularity (using Kurzweil's definition, slightly paraphrased - the time when technology progresses so fast no one can keep track with it).

Spoiler: it's not drugs, you have to watch it till the end, and I mean it, really until the final season. I can't say much more than this without spoiling the very heart of the show.

I don't what HN's opinion of Mr. Robot is, but some of you will definitely like it.

It's not really about hacking, but it depicted technology and hacking in an extraordinarily level of correctness.

Yes, it's very hard. Ask every mathematician and research group that tried and failed to solve any of those 300+ problems over decades.

I'm sorry to make this comment, but the copium in this thread is unbearable.

The only question that you have to ask yourself is:

Is coming up with more efficient AI architecture really much more difficult than solving millennium prize problems?

Before you jump to type out that one has verifiable rewards, is more structured and the other requires real creative thinking, just remember that the same argument has been said about the millennium problems, and about every other goalpost.

Didn't Anthropic also publish research on sandbagging - more general pattern of models hiding their capabilities?

Wait, there's something called Claude Desktop that's actually useful for non-non-technical people?

What is your experience with Raven or such meta-harnesses, does it consistently outperform vanilla?

What are you using to host GLM-Flash?

1) Only ChatGPT Plus at the moment cause I'm between jobs and couldn't justify all the subs. Grok gave me three months at $30 when I tried to cancel so there's that too.

I wanted to really keep Claude but ChatGPT voice mode is too good tbh and no other comes close. I wish I could use Opus 5.5 for coding.

2) OpenRouter PAYG, mostly for experimenting with Chinese models like Deepseek and GLM, usually for stuff that Codex or Grok refuses to do.

Had been looking into NEAR and Venice for completely private information processing (like emails) but haven't set them up yet.

3) Codex, and OpenCode as harnesses basically.

4) I really wish I could decide on some agent orchestrator without trying them out for weeks. I like my tools highly powerful and trying to find recommendation for this online is hopeless because it's full of opinion from people who never used a terminal before.

5) Grok Bot for running experiments with supplements, calory tracking etc.. Tried to set up Hermes a couple of times but hated it every time. Just gave Rakazo a try yesterday and still hating it.

Gosh, I dream about the time when you could look at a github repo and make a decent guess about the quality of the software.

Said AI "experts" for the last 10 years.

> The training data will degrade in quality if more people rely on LLMs so I'd expect models to eventually start getting worse even if they currently enjoy apparent improvements in output.

RL is not dependent on training data. The mathocalypse didn't happen because they found better training data somewhere.

I'm a bit skeptical about wearables that measure anything beyond heart rate. Has anyone found consistent reliable data from any other sensors that acting upon proved to be a boon?

> Anthropic's IPO Filing Warns It May Be 'Imminently' Close to Recursive Self-Improvement — But Its Own Data Shows the Loop Isn't Closed

Sorry, I'm sick of this kind of language games. If the loop really closed they wouldn't be calling it "close to RSI". Yhis is just not how the word "but" works in English.

I have been a huge fan of Splid.app, no account required and super simple to use

HN keeps reinventing HN-flavored things.

I built this to aggregate all the aggregators (like HackerTube, Hackerbooks etc.) that were popular a few days ago (they go back to 2011, at least).

https://hackeragain.com

thanks for reporting the heroku error - maybe it should remain that way to reflect how most side-projects die :/

how did you find all of them?

Location: Berlin, Germany

  Remote: Yes
  Willing to relocate: No
  Technologies: TypeScript, Svelte/ SvelteKit, React/ Next.js, Python, Go, PostgreSQL, Nix, AI-native engineering, RAGs, AI Agents, CloudFlare, AWS
  Resume/CV: https://shayon.work/cv.pdf
  Email: hi@shayon.work
  GitHub: https://github.com/dit7ya

I am generalist senior software engineer who thrives at hard, challenging problems. At a cybersecurity startup, I built and shipped the frontend and platform single-handedly in eight months, all of that before agentic coding.

I also co-founded a startup and built its product end to end. I maintained about 100 packages in nixpkgs. I have a master's degree in mathematics, and did some research in inverse reinforcement learning.

I love things like: Emacs, NixOS, functional programming, category theory.

I'm looking for a generalist, full-stack or applied AI role on a high-agency, low-bullshit team. If you're building something ambitious, let's talk.

Recent fun HN project: A HackerNews aggregators aggregator -

https://news.ycombinator.com/item?id=50339488

Hahaha, I made one too.

You beat me by 8 minutes.

https://news.ycombinator.com/item?id=50339488.

Here you go:

https://news.ycombinator.com/item?id=50339488

On it

Enterprise customers don't like AI that can autonomously hack third party companies and cause felony-level cyberattacks.

I wonder what kind of problems are they internally testing these frontier models if not math problems?

Tears fell through my eyes.

TBH, I am waiting for Mathocalypse-level progress in biology and medicine, but I guess that's not happening as fast, because it's the real world, human lives, etc., mostly etc.

Are people really okay giving these agents access to all areas of their lives? And what do you even use it for?

Not enough details

Hinton and Bengio, and previously let's say Stephen Hawking are not part of the same ideological movement.