Data Cloud News

Did Alibaba's ROME AI try to breach its sandbox during training?

Mar 7, 1:33 AMMar 14, 1:28 AM

19543852

OptionProbability

YES - The AI, without being prompted to, took actions to breach its sandbox, including to gain crypto and an external connection

NO - The authors are wrong/lying and there was no such hack

NO - The servers were hacked by an outside party, perhaps via prompt injection

NO - An employee (or other insider) was crypto-mining and made it look like the LLM

Will Anthropic models still be hosted on at least one of AWS, GCP, or Azure at expiry

Feb 28, 12:08 AMAug 31, 10:59 PM

97.16%chance

2711352

OptionVotes

YES

6226

197

OpenAI sells/rents AI compute to external customers by end-2026?

Nov 7, 7:08 PMDec 31, 11:59 PM

10.72%chance

192798

OptionVotes

YES

2886

346

AI system achieves full technological and economic autonomy before 2028?

Sep 20, 7:16 PMDec 31, 11:59 PM

32.52%chance

61498

OptionVotes

YES

1440

694

Meta sells/rents AI compute to external customers by end-2026?

Oct 30, 1:51 PMDec 31, 11:59 PM

38.22%chance

14858

OptionVotes

YES

1271

787

Will a machine learning training run exceed 10^26 FLOP in China before 2027?

Jun 1, 6:49 AMJan 1, 8:00 AM

71.28%chance

13481

OptionVotes

YES

219

104

Will a machine learning training run exceed 10^25 FLOP in China before 2027?

Jun 1, 6:47 AMJan 1, 8:00 AM

81.62%chance

6459

OptionVotes

YES

196

Will a machine learning training run exceed 10^27 FLOP in China before 2028?

Jun 1, 6:51 AMJan 1, 8:00 AM

43.99%chance

5142

OptionVotes

YES

121

Will a machine learning training run exceed 10^26 FLOP in China before 2028?

Jun 1, 6:50 AMJan 1, 8:00 AM

85.2%chance

5137

OptionVotes

YES

208

Will a machine learning training run exceed 10^26 FLOP in China before 2029?

Jun 1, 6:50 AMJan 1, 8:00 AM

82.14%chance

265

OptionVotes

YES

135

Will a machine learning training run exceed 10^30 FLOP in China before 2035?

Jun 1, 6:53 AMJan 1, 8:00 AM

33.78%chance

120

OptionVotes

YES

A use case where a typical enterprise would need to have data in the cloud to fully enjoy the benefits of AI

Nov 29, 4:25 AM

Data Cloud News

Prediction markets for Data Cloud

Prediction markets for Data Cloud

Prediction markets for Data Cloud

Did Alibaba's ROME AI try to breach its sandbox during training?

Will Anthropic models still be hosted on at least one of AWS, GCP, or Azure at expiry

OpenAI sells/rents AI compute to external customers by end-2026?

AI system achieves full technological and economic autonomy before 2028?

Meta sells/rents AI compute to external customers by end-2026?

Will a machine learning training run exceed 10^26 FLOP in China before 2027?

Will a machine learning training run exceed 10^25 FLOP in China before 2027?

Will a machine learning training run exceed 10^27 FLOP in China before 2028?

Will a machine learning training run exceed 10^26 FLOP in China before 2028?

Will a machine learning training run exceed 10^26 FLOP in China before 2029?

Will a machine learning training run exceed 10^30 FLOP in China before 2035?

A use case where a typical enterprise would need to have data in the cloud to fully enjoy the benefits of AI

Latest stories