Microsoft CEO Says Tech Giants Battling For Content To Build AI

Microsoft chief executive Satya Nadella said Monday tech giants were competing for vast troves of content needed to train artificial intelligence, and complained Google was locking up content with expensive and exclusive deals with publishers. From a report: Testifying in a landmark U.S. trial against its rival Google, the first major antitrust case brought by the U.S. since it sued Microsoft in 1998, Nadella testified the tech giants' efforts to build content libraries to train their large language models "reminds me of the early phases of distribution deals." Distribution agreements are at the core of the U.S. Justice Department's antitrust fight against Google. The government says that Google, with some 90% of the search market, illegally pays $10 billion annually to smartphone makers like Apple and wireless carriers like AT&T and others to be the default search engine on their devices.

The clout in search makes Google a heavy hitter in the lucrative advertising market, boosting its profits. Nadella said building artificial intelligence took computing power, or servers, and data to train the software. On servers, he said: "No problem, we are happy to put in the dollars." But without naming Google, he said it was "problematic" if other companies locked up exclusive deals with big content makers. "When I am meeting with publishers now, they say Google's going to write this check and it's exclusive and you have to match it," he said.

Not for me.

[midifarm]

Just not interested. Not afraid to be the Luddite in the room.

Re:

[VampireByte]

Then you aren't a Luddite. The actual Luddites were very interested.

Re:

[Tony Isaac]

Just like the original Luddites, you have a choice. You can learn about and adapt to the new technology, or you can rage against it and be overcome by the tsunami that will spare no one.

LLMs are not a passing fad. They provide real productivity benefits for many kinds of people in many kinds of professions. Coding is an obvious example. These days, we find ourselves writing code in multiple languages in any given project, some of which we know better than others. GitHub Copilot and ChatGPT are pretty great

Oh great.

[Black Parrot]

Now bookstores will be filled with knock-offs written as AI fodder.

If only there was this domain of content.

[xack]

That was open to the public. Oh, you restricted it because of a cartoon mouse? Right.

When monopolies battle monopolies

[Tablizer]

I get the popcorn & booze out and watch them catfight and cry.

Maybe some hope on copyright

[superzerg]

If big players like Google and Microsoft start to feel annoyed by the present crazy copyright terms, there might be some hope for change

Google providing deals?

[Somefriend]

What authorize google to distribute license to content they don't own exactly? Data needs to be negotiated with entities that actualy produce them, google is barely a distribution platform. They have no rights to sell other's content.

Re:

[codebase7]

Google / Microsoft / OpenAI / etc. aren't distributing the copyrighted content itself, rather they are selling the product produced by having knowledge of the copyrighted content.

I.e. Because Google / Microsoft / OpenAI / etc. are selling the output produced by a virtual "brain" that contains copyrighted content, to be legally safe Google / Microsoft / OpenAI / etc. must license the copyrighted content said "brain" contains. If TFS is to be believed, then apparently Google has found it advantageous to inc