• Jensen Huang said people programming AI is similar to the way "you program a person."
• Speaking at London Tech Week, the Nvidia CEO said all anyone had to do to program AI was "just ask nicely."
• He called AI "the great equalizer, " allowing anyone to program computers using plain language.

Nvidia CEO Jensen Huang has said that programming AI is similar to "the way you program a person" — and that "human" is the new coders' language.

"The thing that's really, really quite amazing is the way you program an AI is like the way you program a person," Huang told London Tech Week on Monday.

Huang shared an example, saying, "You say, 'You are an incredible poet. You are deeply steeped in Shakespeare, and I would like you to write a poem to describe today's keynote.' Without very much effort, this AI would help you generate such a wonderful poem.

"And when it answers, you could say, 'I feel like you could do even better.' And it will go off and think about it and it will come back and say, 'In fact, I can do better.' And it does do a better job."

Huang said that in the past, "technology was hard to use" and that to access computer science, "we had to learn programming languages, architect systems, and design very complicated computers.

"But now, all of a sudden, there's a new programming language. This new programming language is called human."

"Most people don't know C++, very few people know Python, and everybody, as you know, knows human."

Huang called AI "the great equalizer" for making technology accessible to everyone and called the shift "transformative.

"This way of interacting with computers, I think, is something that almost anybody can do," he said.

"The way you program a computer today is to ask the computer to do something for you, even write a program, generate images, write a poem — just ask it nicely," Huang added.

At the World Government Summit in Dubai last year, Huang suggested the tech sector should focus less on coding and more on using AI as a tool across fields like farming, biology, and education.

"It is our job to create computing technology such that nobody has to program. And that the programming language is human, everybody in the world is now a programmer. This is the miracle of artificial intelligence," Huang said at the time.

• Google CEO Sundar Pichai said the company is tracking how AI makes its engineers more productive.
• During the "Lex Fridman Podcast," Pichai estimated a 10% increase in engineering capacity.
• Separately, Google and Microsoft have publicly shared how much of their code is being generated by AI.

Google is tracking how AI is making its engineers more productive — and has developed a specific way to measure it.

Speaking on an episode of the "Lex Fridman Podcast" that aired last week, Google CEO Sundar Pichai said that the company was looking closely at how artificial intelligence was boosting productivity among its software developers.

"The most important metric, and we carefully measure it, is how much has our engineering velocity increased as a company due to AI?" he said. The company estimates that it's so far seen a 10% boost, Pichai said.

A Google spokesperson clarified to Business Insider that the company tracks this by measuring the increase in engineering capacity created, in hours per week, from the use of AI-powered tools.

Put simply, it's a measurement of how much extra time engineers are getting back thanks to AI.

Whether Google expects that 10% number to keep increasing, Pichai didn't say. However, he said he expects agentic capabilities — where AI can take actions and make decisions more autonomously — will unlock the "next big wave".

Google has its own internal tools to help engineers code. Last year, the company launched an internal coding copilot named "Goose," trained on 25 years of Google's technical history, Business Insider previously reported.

While AI Pichai said during the podcast that Google plans to hire more engineers next year. "The opportunity space of what we can do is expanding too," he said, adding that he hopes AI removes some of the grunt work and frees up time for more enjoyable aspects of engineering.

Separately, the company is tracking the amount of code that is being generated by AI within Google's walls — a number that is apparently increasing.

Pichai said during Alphabet's most recent earnings call that more than 30% of the company's new code is generated by AI, up from an estimated 25% in October.

Google isn't the only one. Speaking at London Tech Week on Monday, Microsoft UK CEO Darren Hardman said its GitHub Copilot coding assistant is now writing 40% of code at the company, "enabling us to launch more products in the last 12 months than we did in the previous three years."

He added: "It isn't just about speed."

In April, Meta CEO Mark Zuckerberg predicted AI could handle half of Meta's developer work within a year.

Additional reporting by Effie Webb.

Have something to share? Contact this reporter via email at [email protected] or Signal at 628-228-1836. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

We've been expecting it for a while, and now it's here: OpenAI has introduced an agentic coding tool called Codex in research preview. The tool is meant to allow experienced developers to delegate rote and relatively simple programming tasks to an AI agent that will generate production-ready code and show its work along the way.

Codex is a unique interface (not to be confused with the Codex CLI tool introduced by OpenAI last month) that can be reached from the side bar in the ChatGPT web app. Users enter a prompt and then click either "code" to have it begin producing code, or "ask" to have it answer questions and advise.

Whenever it's given a task, that task is performed in a distinct container that is preloaded with the user's codebase and is meant to accurately reflect their development environment.

Read full article

Comments

There are few areas where AI has seen more robust deployment than the field of software development. From "vibe" coding to GitHub Copilot to startups building quick-and-dirty applications with support from LLMs, AI is already deeply integrated.

However, those claiming we're mere months away from AI agents replacing most programmers should adjust their expectations because models aren't good enough at the debugging part, and debugging occupies most of a developer's time. That's the suggestion of Microsoft Research, which built a new tool called debug-gym to test and improve how AI models can debug software.

Debug-gym (available on GitHub and detailed in a blog post) is an environment that allows AI models to try and debug any existing code repository with access to debugging tools that aren't historically part of the process for these models. Microsoft found that without this approach, models are quite notably bad at debugging tasks. With the approach, they're better but still a far cry from what an experienced human developer can do.

Read full article

Comments