Platformonomics TGIF #68: October 11, 2024

By

Platformonomics TGIF is a weekly roll-up of links, comments on those links, and perhaps a little too much tugging on my favorite threads.

Another week of our favorite threads unspooling…

News

Antitrust Incoherence: Google Breakup Rumors

We’re going through another round of Google remedy rumors. Some are legal posturing, some are competitors helpfully spitballing ideas that just happen to maximally advantage them at Google’s expense, and some are the Tourette’s-like incantation of Hipster Antitrust‘s (movement slogan: Details aren’t important) credo: “Break them up!”

I’m still trying to understand how tying the browser to a monopoly asset has gone from an apparently big deal antitrust violation unto itself to merely a potential remedy. But that desire for a coherent antitrust doctrine does complicate my life.

I’ve mentioned this before, but having been at Microsoft during its antitrust agonies, it is hard to understate the impact on morale at Google. It wears on employees to see those kind of headlines every day. And Microsoft was only losing one antitrust case.

Media Self Harm: Lobbying vs. Innovating

Per The Information:

Publishers are happy that the Department of Justice sued Google over its control of online display advertising. But some would have preferred to have seen the case brought five or six years ago—when the gap between what publishers and the tech giants generate from online ads was not as wide as it is today.

Publishing executives say they might have had a chance to build the ad tools that could have helped them compete for ad dollars if Google’s control of buying and selling display ads on the open web had been loosened back then.

The juxtaposition of the media pushing revisionist history about their Internet failures while committing the exact same error with AI today is awfully rich.

We’re supposed to believe that if antitrust action against Google’s ad business had happened just a little earlier, the publishers would have innovated and invested to build their own digital ad platforms. Yet at the very same time these media brunchlords are missing in action when it comes to innovating and investing with AI.

The media companies have taken some payoffs from frontier model companies (or are litigating in hopes of getting a little more money), but they are dedicating far more internal effort to lobbying governments to make Big Tech pay them for media’s existential digital incompetence in the last era (another variant of the Time Travel Theory of Antitrust).

Once again, as we did with search and social networks, we have to ask why media is not leading the innovation with large language models? They certainly are among the most affected. (And, yes, I do know the answer to that question and will admit it seems preposterous to ask).

Starting to think my nuclear power cheerleading work is done…

And, perhaps the clearest confirmation that nuclear power is back, aspiring tech company/cookie colossus/puzzle purveyor The New York Times is in full hand-wringing mode about the budding nuclear renaissance (though they do uncharacteristically admit they don’t know what they’re talking about in the next to last paragraph of that piece).

…except in Taiwan

TSMC, the world’s leading (and leading edge) chip fab, is 8% of Taiwanese electricity consumption, and could go to 24% by 2030.

Energy suicide is a terrible way to go (as the Europeans are showing).

Obligatory NotebookLM Podcast: Antitrust Edition

Here’s the Google AI-generated podcast for the antitrust part of my A New Antitrust Doctrine post. It may be better than the median podcast, though needs to learn to pronounce “Brandeisian”.

Blog Roll Addition: Tech Can Be Better

John Ludwig reflects on on the past and future of technology in his new-ish blog Tech Can Be Better, with a particularly relevant focus on why automakers can’t do software to save their lives (which is a question of life and death for them).

2 responses

  1. I was pretty impressed (granted everyone is by demos) by Palantir’s customer support agent that intelligently used data (properly organized in a perfect implementation) to give better and faster answers than having to go through the usual chain of initiating a phone call to requesting a live agent to then requesting escalation to get something solved. I’ve been saying forever that if you can truly automate support based on a massive database of issue-and-answer that’s a whole new industry. Every platform has support, which usually is suboptimal because you lose senior support people to sales or a higher incentive. If someone gets it right, that’s the next SNOW or Salesforce. I doubt the incumbents win this.

  2. Charles Fitzgerald Avatar

    Agree support agents will be big. Lots of contenders. And it won’t be Salesforce (they’ll end up having to buy someone).

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Get Updates By Email

Discover more from Platformonomics

Subscribe now to keep reading and get access to the full archive.

Continue reading