In poker, the value of the card in your hand is often your rivals feel that you can catch. As long as you can explain to others, you do not need royal flush.
Openai CEO Sam Altman knows this well, played a large -scale poker during his student days. After the astronomical success of its normal artificial intelligence tool chat, the company has assured many backers that it holds all aces, telling the world that the scale is the key to progress and betting on it will receive big awards.
On 21 January, Altman announced Stargate, which plans a $ 500 billion to create a huge data center for future AI models. As he said in an interview in 2023: “It is completely disappointing to compete with us.”
But the Chinese AI company Deepsek now calls its bluff. It sent a shock waves through the Silicon Valley with the release of the AI model in the last two weeks, which are clearly capable of Openai’s best, but at a fraction of cost and computational power (see “Deepsek energy demands. Shows a way to reduce? ”. Keep it or need such huge resources to do so.
Deepsek has punctured the idea that American companies take a secret recipe for the manufacture of AI.
For those concerned about the accumulation of power in Silicon Valley, the arrival of the competition is welcome, but the model of Deepsek brings its own concerns. For one thing, its north is close to the party line of the Chinese government, and also Censor in real timeSecurity researchers have also warned that this There is a lack of strong railings against unfair use,
Nevertheless, its arrival on the scene suggests that there are huge innovations in liberal AI to come yet. In addition, cheap models that require less computational power must open the door for completely new applications for technology, which can make it cheap for more people and less harmful for planets. With more players around the table, bets may not be high.
Subject:
(Tagstotransite) Environment (T) Artificial Intelligence (T) Technology
