Outcome-Based Reinforcement Learning to Predict the Future

96•bturtel•1d ago

Comments

ctoth•1d ago

Do you want paperclips? Because this is how you get paperclips!

Eliminate all agents, all sources of change, all complexity - anything that could introduce unpredictability, and it suddenly becomes far easier to predict the future, no?

JoshTriplett•1d ago

> Do you want paperclips? Because this is how you get paperclips!

Don't^W worry, there are many other ways of getting paperclips, and we're doing all of them.

sitkack•1d ago

Even explaining how not to get paper clips, gets you paper clips when you can invert the loss function. Paper clips for everyone!

vlovich123•1d ago

I don't know. Paperclips are awful useful. Would it be so bad to build more of them?

throwaway71271•1d ago

https://www.decisionproblem.com/paperclips/index2.html go ahead :)

Ygg2•23h ago

That's all fun and games until paperclip maximizers starts looking at your blood as source of iron.

valine•1d ago

So instead of next token prediction its next event prediction. At some point this just loops around and we're back to teaching models to predict the next token in the sequence.

lumost•1d ago

Tokens are an awfully convenient way to describe an event.

phyalow•23h ago

Tokens are just discretized state representations.

ww520•1d ago

It’s the next state. So instead of spitting out words, it will spit out a whole movie, or a sequence of world states in a game or simulation.

amelius•22h ago

Why would you use RL if you're not going to control the environment, but just predict it?

TOMDM•19h ago

Because they're training a predictor, not an agent?

jldugger•22h ago

From the abstract

> A simple trading rule turns this calibration edge into $127 of hypothetical profit versus $92 for o1 (p = 0.037).

I'm lazy: is this hypothetical shooting fish in a barrel, or is it a real edge?

nyrikki•21h ago

Note the 'hypothetical profit' part , I know of several groups looking for opportunities to skim off LLM traders, leveraging its limited sensitivity, expressiveness, and the loss of tail data.

Predictive AI is problematic no matter what tool you use. Great at demoware that doesn't deliver.

I am sure there are use cases, but it would be augmentation, not a reliable approach by itself.

garbagecoder•1h ago

"a couple of wavy lines"

bzzzzz "sorry this isn't your lucky day"

Show HN: I rewrote my Mac Electron app in Rust

Compiler Explorer and the promise of URLs that last forever

Compiling a Neural Net to C for a 1,744× speedup

What does "Undecidable" mean, anyway

A toy RTOS inside Super Mario Bros. using emulator save states

Ice Cream Replaced Booze in the US Navy

Visualize and debug Rust programs with a new lens

Deepseek R1-0528

Show HN: Tesseral – Open-Source Auth

LLM Codegen go Brrr – Parallelization with Git Worktrees and Tmux

The Blowtorch Theory: A new model for structure formation in the universe

Collatz's Ant and Similarity of Landscapes

GoGoGrandparent (YC S16) is hiring Back end Engineers

Mathematical Fiction

As a developer, my most important tools are a pen and a notebook

Launch HN: MindFort (YC X25) – AI agents for continuous pentesting

Getting a Cease and Desist from Waffle House

xAI to pay telegram $300M to integrate Grok into the chat app

The mysterious Gobi wall uncovered

Show HN: Loodio 2 – A Simple Rechargable Bathroom Privacy Device

De-anonymization attacks against the privacy coin XMR

Show HN: My LLM CLI tool can run tools now, from Python code or plugins

Square Theory

Implementing complex numbers and FFT with just datatypes (2023)

Building interactive web pages with Guile Hoot

Show HN: Wetlands – a lightweight Python library for managing Conda environments

Show HN: AutoThink – Boosts local LLM performance with adaptive reasoning

A thought on JavaScript "proof of work" anti-scraper systems

Homo erectus from the seabed, new archaeological discoveries in Indonesia

The Level Design Book