A prediction market on Manifold ( asks whether DeepSeek lied about the GPU compute budget used to train v3.
A prediction market on Manifold ( asks whether DeepSeek lied about the GPU compute budget used to train v3. As of June 21, 2026, the market gives a 'yes' probability of only 4.01%, suggesting low community belief in deception. However, the mere existence of this question signals ongoing trust concerns in DeepSeek's reported efficiency. If true, it would undermine DeepSeek's claims of cost-efficient training, potentially eroding developer trust and investor confidence. The low probability limits conviction, but the signal is aligned with collection interest in developer ecosystem trust.
signal brief
A prediction market on Manifold (https://manifold.markets/AndrewPeterson960c/did-deepseek-lie-about-the-number-o) asks whether DeepSeek lied about the GPU compute budget used to train v3. As of June 21, 2026, the market gives a 'yes' probability of only 4.01%, suggesting low community belief in deception. However, the mere existence of this question signals ongoing trust concerns in DeepSeek's reported efficiency. If true, it would undermine DeepSeek's claims of cost-efficient training, potentially eroding developer trust and investor confidence. The low probability limits conviction, but the signal is aligned with collection interest in developer ecosystem trust.
evidence
Decision support, not stock advice. This signal is research with cited evidence — not a recommendation to buy, sell, or hold any security.