compute
US /kəmˈpjut/
・UK /kəm'pju:t/
Video subtitles
Google I/O '24 in under 10 minutes

- Trillium delivers a 4.7x improvement in compute performance per chip over the previous generation.
Trillium delivers a 4.7x improvement in compute performance per chip over the previous generation.
Inside the Making of an AI Chip | WSJ Tech Behind

- These are the compute elements or the components that actually perform the computation.
These are the compute elements or the components that actually perform the computation.
- That compute die then gets integrated into a package, and that's what people typically think about when they think about the chip.
That compute die then gets integrated into a package, and that's what people typically think about when they think about the chip.
A recap of ChatGPT | tech news

- for its compute power to train and run their models
for its compute power to train and run their models
Dividing numbers: intro to long division | 4th grade | Khan Academy

- So we're going to try to compute what 96 divided by 4 is.
So we're going to try to compute what 96 divided by 4 is.
Priority Seats and Invisible Disabilities | Fashioneyesta

- And there are those people who don't even seem to compute that disabled people are entitled to sit in priority seats as well.
And there are those people who don't even seem to compute that disabled people are entitled to sit in priority seats as well.
- But there are also those who don't understand that and even when you do explain it to them they don't seem to compute what you're saying.
But there are also those who don't understand that and even when you do explain it to them they don't seem to compute what you're saying.
DeepSeek R1 - The Chinese AI "Side Project" That Shocked the Entire Industry!

- On January 20th, 2025, a small Chinese research firm called DeepSeek released DeepSeek R1, a completely open source, open weights, AI model that has the ability to think, also known as test time compute, that is directly competitive, if not slightly better than the O1 model by OpenAI that cost hundreds of millions of dollars to train.
On January 20th, 2025, a small Chinese research firm called DeepSeek released DeepSeek R1, a completely open source, open weights, AI model that has the ability to think, also known as test time compute, that is directly competitive, if not slightly better than the O1 model by OpenAI that cost hundreds of millions of dollars to train.
- We've been talking on this channel a lot about test time compute.
We've been talking on this channel a lot about test time compute.
DeepSeek是在“抄袭”美国吗?它的出现对中美AI竞争的影响 DeepSeek | LLM |Open AI | 中国 |美国 |人工智能竞争 |开源模型 20250126金融汪

- And they did roughly $5 million in total compute budget.
And they did roughly $5 million in total compute budget.
- And the amount of compute DeepSeq has access to is roughly similar to what PhD students in the US have access to.
And the amount of compute DeepSeq has access to is roughly similar to what PhD students in the US have access to.
A shocking Chinese AI advancement called DeepSeek is sending US stocks plunging

- And the compute cost is a fraction.
And the compute cost is a fraction.
- The question is, going forward, is algorithmic efficiency, this ability to run smaller models more efficiently, going to be a better path forward than the brute force, massive compute models that we've used so far at OpenAI and Anthropic and at Google Gemini and at Metas Llama?
The question is, going forward, is algorithmic efficiency, this ability to run smaller models more efficiently, going to be a better path forward than the brute force, massive compute models that we've used so far at OpenAI and Anthropic and at Google Gemini and at Metas Llama?
Introducing Dagster+

- You can see things like the latest status of your asset, including the data quality testing configured on the asset, its description, who owns the asset, what groups and tags are associated with it, as well as where the asset is located, its compute details, and the associated resources and compute kind.
You can see things like the latest status of your asset, including the data quality testing configured on the asset, its description, who owns the asset, what groups and tags are associated with it, as well as where the asset is located, its compute details, and the associated resources and compute kind.
- When automatically ingested from DBT or derived from user-provided metadata, Dagster can compute the relationship of individual columns and represent them on the page similar to our asset lineage graph.
When automatically ingested from DBT or derived from user-provided metadata, Dagster can compute the relationship of individual columns and represent them on the page similar to our asset lineage graph.
Casually Explained: Bitcoin

- While this has begun changing a lot in the last few years, after so many people were losing all their money repeatedly, there is a lot of environmental criticism brought towards crypto because the process of mining new Bitcoin is extremely energy intensive as the process is done by brute forcing math problems with a huge amount of compute.
While this has begun changing a lot in the last few years, after so many people were losing all their money repeatedly, there is a lot of environmental criticism brought towards crypto because the process of mining new Bitcoin is extremely energy intensive as the process is done by brute forcing math problems with a huge amount of compute.