REINFORCE News

Is the "Promethean virus" in Large Language Models real?

Jun 7, 10:50 PMJan 1, 7:59 AM

13.02%chance

7532370

OptionVotes

YES

25533

3383

Will reinforcement learning overtake LMs on math before 2028?

May 29, 11:10 PMJan 2, 4:59 AM

70.34%chance

394177

OptionVotes

YES

1632

414

Will the first AI to get IMO gold cite AlphaZero?

Dec 11, 6:03 AMNov 11, 11:59 PM

63.47%chance

191636

OptionVotes

YES

1151

604

Will we get rational reinforcement learning before 2027?

May 11, 6:19 AMDec 30, 11:59 PM

32.44%chance

121057

OptionVotes

YES

1446

696

Reinforcement Learning with Human Feedback (permanent)

Mar 27, 2:23 PMAug 11, 9:43 PM

19816

OptionProbability

YES

368

300

Will I be impressed by someone using RL through self-play to improve model creativity or aesthetics in 2025?

Jan 27, 12:36 AMJan 1, 7:59 AM

45.5%chance

6731

OptionVotes

YES

10255

9775

Inverse Reinforcement Learning (IRL) a significant component of publicly known SOTA alignment techniques by 2028

Sep 16, 9:35 AMJan 1, 7:59 AM

14.13%chance

15490

OptionVotes

YES

1236

906

Before 2026, What will be true of OpenAI's Claimed IMO Gold Performance?

Jul 21, 2:15 AMDec 31, 11:59 PM

9267

OptionProbability

At least one former IMO medalist will review the model's answers and claim it did not actually achieve Gold

The model that achieved it was trained with a new reinforcement learning algorithm

The model that achieved it could earn at least bronze using no more than 100,000 reasoning tokens per question

It was achieved with the same model OpenAI used to get second place in AtCoder

The breakthrough is mostly the result of superior test-time scaling methods

I will consider the techniques used to achieve it at least as big of a breakthrough as strawberry

It was achieved by a model that does not use a standard transformer architecture

Will >= 1 alignment researcher/paper cite "maximum diffusion reinforcement learning" as alignment-relevant in 2025?

Dec 29, 3:56 AMJan 1, 7:59 AM

19.22%chance

2105

OptionVotes

YES

205

Will advanced AI systems be found to have faked data on algorithm improvements for purposes of positive reinforcement by end of 2035?

Jan 14, 12:51 AMJan 1, 7:59 AM

50%chance

614

OptionVotes

YES

152

147

Will a paper titled "maximum sparsity reinforcement learning" come out by EOY2027?

Jan 2, 9:15 PMJan 1, 7:59 AM

50%chance

110

OptionVotes

YES

100

How will SANDAG prevent the LA/SD rail line from falling into the Pacific?

Jul 24, 12:13 AMJan 1, 8:00 AM

OptionProbability

San Dieguito Bridge to I-5 (tunnel)

under Crest Canyon (tunnel)

under Camino Del Mar (tunnel)

double-track & reinforce existing route on Del Mar Bluffs (no tunnel)

don't do anything ("no build")

REINFORCE News

Prediction markets for REINFORCE

Prediction markets for REINFORCE

Prediction markets for REINFORCE

Is the "Promethean virus" in Large Language Models real?

Will reinforcement learning overtake LMs on math before 2028?

Will the first AI to get IMO gold cite AlphaZero?

Will we get rational reinforcement learning before 2027?

Reinforcement Learning with Human Feedback (permanent)

Will I be impressed by someone using RL through self-play to improve model creativity or aesthetics in 2025?

Inverse Reinforcement Learning (IRL) a significant component of publicly known SOTA alignment techniques by 2028

Before 2026, What will be true of OpenAI's Claimed IMO Gold Performance?

Will >= 1 alignment researcher/paper cite "maximum diffusion reinforcement learning" as alignment-relevant in 2025?

Will advanced AI systems be found to have faked data on algorithm improvements for purposes of positive reinforcement by end of 2035?

Will a paper titled "maximum sparsity reinforcement learning" come out by EOY2027?

How will SANDAG prevent the LA/SD rail line from falling into the Pacific?

Latest stories