Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
“There are very few companies that are not affected by the AI on Monday on Monday. Mobile World Congress In Barcelona. He rejected the idea that open source models could reduce calculation demand and when asked when European agencies would change their generator AI strategies in the light of geo -political tensions in the United States
Sharma said at the Startup Conference at the Startup Conference that Amazon is now deploying AMazon Web Services – Amazon’s Cloud Computing Division – Alexa Consumer Product in its warehouse through its own foundational models.
“We now have something like three-quarters of a million robots and they are doing everything from picking up things until they run themselves in the warehouse. The Alexa product is probably the most widely deployed in existence … there is no part of Amazon that is unclear by the generator AI. “
December, AWS Declaration A new suit of four text-renowned models, a family of multimodal generator AI models called it Nova.
Sharma said that these models were tested against public criteria: “It is quite clear that there is a huge variation in use. There is no place. There are some places where your video generation is needed … and other places such as Alexa, where you ask it to do a specific work and the response needs to be very, very quick, and it needs to be very fast.”
However, he said that calculations with small open source models were less likely to be reduced: “When you start applying it to different situations, you only need more and more intelligence,” he said.
Amazon has also launched “Bedrock“A service agencies and startups in the AWS are aiming to mix and match different foundational models – including – including DIPSECThe It enables the users to switch seamlessly to the models, he said.
Amazon is also creating a huge AI Computer Cluster in its trainium 2 chips in an anthropic partner, where it has invested $ 8 billion. Mean Have been released Its latest flagship AI Model, Grock 3, uses a huge data center in Memphis that has about 200,000 GPUs.
Asked about this level of counting resources, Sharma said: “My personal opinion is that the counting will be a part of the long -standing conversation.”
He did not think that Amazon was under pressure from the snow flashes of open source models recently derived from China: “I wouldn’t describe it like this,” he said. On the contrary, the Amazon AWS felt comfortable deploying dipsek and other models in the AWS, he advised. Sharma said, “We believe in a company that loves … we are open to any trend and technology from the customer’s point of view,” Sharma said.
When the Open AI ChatzP was launched by the end of 2022, did he think Amazon was caught in Napping?
“No, I think I don’t agree with that line of thought,” he said. “Amazon has been working on AI for about 25 years. If you see something like Alexa, there are 20 different AI models that are running in Alexea … We had a few billion parameters that already existed for language. We have been watching it for quite some time. “
In the recent issue Debate The surrounding Trump and Zensky and subsequently Strain Regarding US relations with many European countries, he thinks that European companies can see somewhere else for the future geni resources?
Sharma acknowledged that this problem was “outer” of its “efficiency” and the consequences “is very difficult for me to predict me …” but he somewhat diplomatically indicated that some companies could adjust their strategy. “What I say is to say that technical innovation responds to the encouragement,” he said.
