‘Sandbox first’: Andrew Ng’s blueprint for accelerating enterprise AI innovation

[ad_1]

Join a reliable event by enterprise leaders in about two decades. VB Transform, Real Enterprise AI strategy brings together people who build. Learn more


Enterprises may be concerned when facilities are produced, but these projects will prevent these projects with guard.

Founder, founder Content AI One of the most prominent people in the EU development also stressed the importance of observation and guards in the development of the AI ​​during firefighting Turn a vb today. However, it added that these should not come to innovation and growth.

NG proposed to quickly find prototype projects in the prototype projects and proven to be established in prototype projects and investigate these applications after proven to work. This can be reflected in the enterprises to apply AI.

>> – siSee all transforms 2025 coverage of our entire transform<

“There is an important role for observation, security and guards,” Ng said. “Then it tends to open them clearly, because it should be stopped by one thing of large enterprises, and the five vice president must be signed.”

Added that large enterprises “Unable to send a slightly casual innovation team in something that harms the brand”, but this can also prevent innovation.

Instead, NG, Sandboxes, developer teams said that “limited personal information with limited personal information” is a way to make a way. Sandboxes allow the organization to invest in only working projects and then add them to technology to include responsible, including observation tools and guardians.

Innovation Sandbox is not rare for businesses, For AI agents, especially. Gumbox Allow for innovation They do not want to go to the public without touching any sensitive information within the enterprise. However, they also allow teams to be as creative as possible to test their views.

Observation of observation is rapidly becoming a key subject as many AI applications and agents. Righteousness Recently updated for Agent Library, AgentForce 3 Provide advanced visibility Additional support for agent performance and interaction standards like MCP.

Speed ​​and low pilot costs

NG is given hand in hand for speed and innovation and enterprises should not be afraid.

“Imagine a roller coaster, but this is a roller coaster, but what happened in the last year, the roller coastal ship took a lot of speed and it’s really exciting because it is progressing.” “I now feel like the world is now in a roll coaster that moves very fast now and is great.”

NG, a factor that contributes to this speed is now available tools Developers work and use frequentlyIt points as coding agents Wind and GitHub Copilot We have reduced development time, “, three months and six engineers who have received me.”

These coding agent platforms and other means that help developers move faster, the price of pilot projects.

“I do not feel that an evidence of a concept is so low that a large number of POCS (the evidence of concept) is bad.”

An obstacle

Although there is an obstacle, you can find talent. The NG admitted that there are AI companies to hire model engineers from up to $ 10 million salary range, but the price for price engineers is not high.

“One of the biggest challenges for many enterprises is talent,” he said. “The good news for the companies seeking engineers have no place in the cost of $ 5 million,” he said.

Although there is a problem, it is still lack of talent to live there in building AI projects for businesses. So NG returns to the first solution: Let them test with sand boxes and make this experience.


[ad_2]
Source link

Leave a Reply

Your email address will not be published. Required fields are marked *