A Federal Judge Just Ruled Google an Illegal Monopoly

Hey SaaS Sentinel reader. Welcome back. This week, AI is training itself on its own data, Google is being declared a monopoly by a federal judge, and 50% of searches are predicted to take place off of search engines next year. Here’s what else is happening in SaaS this week.

• Google's share of the search ad market is projected to dip below 50% next year

• AI researchers are turning to synthetic data to train their models due to data scarcity

• A federal judge ruled Google an illegal monopoly, and the DOJ is looking for fixes

Google's Share of the Search Ad Market Could Drop below 50% for the First Time in a Decade

For over a decade, Google has enjoyed a search advertising market share topping 50% in the US. But that long reign at the top may soon come to an end thanks to the rise of AI search engines like ChatGPT.

According to projections by eMarketer, Google's share of the lucrative search ad market could dip below the 50% threshold next year for the first time in over a decade. And while it’s still the leading search engine by far, Google is facing rapidly growing competition from AI upstarts leveraging large language models that can understand queries and provide results in conversational language.

ChatGPT has seen explosive growth since first launching in November 2022, quickly amassing millions of users. Its early success has pushed Google to rush its own AI-powered search features into the market, including AI-generated responses from its own model, Gemini, atop search results. But Google's efforts to play catchup may be too little too late.

Other AI search engines like Perplexity AI are also gaining serious momentum. Perplexity fielded over 300 million searches in September alone and now has major advertisers lining up for when it launches paid placements. The AI search platform is even allowing its sponsors to provide pre-approved answers for follow-up questions on search results.

Only time will tell whether Google can leverage its own AI innovations to keep pace with the new crop of rivals aiming to end its decade-plus run as the top search advertising dog. But its competition is clearly here to stay.

Read more.

AI Is Training Itself on Its Own Data – Here’s Why It Matters

With quality data increasingly scarce and expensive, AI researchers are turning to synthetic data as a silver bullet solution. Synthetic data is AI-generated training data that promises limitless volume without the costs and biases of human-created data labels. So yes — we’ve come full circle and AI is now using its own data to train itself. These are wild times we’re living in.

Major AI labs like Anthropic, Meta, and OpenAI are already using synthetic data to train flagship natural language models like Claude and Llama. The market for synthetic data generation could be worth over $2 billion by 2030. By sourcing training data from AI systems rather than people, researchers hope to expand the capabilities of generative models faster than real-world data constraints allow.

But there’s a big catch — synthetic data carries several risks if it's not thoroughly vetted by, you know, humans. This is because AI models inevitably perpetuate the flaws and biases present in their own training data. And without careful monitoring, compounding errors can trigger a downward spiral known as a "model collapse," where AI systems spew increasingly nonsensical outputs.

Fortunately, AI researchers say that synthetic data pipelines cannot be fully automated — their outputs need regular inspection, filtering, and augmentation with real-world data to catch issues. Striking the right balance will be key. If the AI community rushes to synthetic data too aggressively, we may end up trading data quality for quantity. 

Read more.

US Considers Asking Courts to Break up Google as It Weighs Remedies in an Antitrust Case

Google is facing a day of reckoning after a federal judge ruled its search engine an illegal monopoly. The Department of Justice is now exploring remedies, including potentially breaking up the tech titan.

In an August decision, Judge Amit Mehta stated Google has long squashed competition by controlling key distribution channels. After a trial scrutinizing deals that pay companies like Apple billions to make Google the default search option, he has ordered DOJ to explore fixes.

DOJ remedies under consideration range from restricting Google's use of data for AI search features to forcing structural splits of products like Chrome and Android. While stopping short of explicitly pursuing a breakup, DOJ calls Google’s search defaults the “starting point” for addressing anticompetitive conduct.

Google warns against “overreach” stifling innovation and says splitting up integrated products would destroy their business models. But with appeals waiting until the remedy phase concludes, Google is stuck playing defense for now.

And the pressure is ramping up both in the US and abroad. Europe is also threatening a potential breakup of Google’s digital ad business. On top of that, Google must open its Android app store to rivals after a separate monopoly ruling.

After years of scrutiny, regulators may finally force seismic changes onto Google’s kingdom. If defaults and data sharing are fair game, even a limited remedy could still rock the foundations of the search empire. And if DOJ ultimately pushes for structural separation, we may one day look back on this as the beginning of the end of Google as we know it.

Read more.

Parting Thoughts

Well, that’s the tech news for this week. Hit reply and let us know—did you learn something from today’s newsletter?

Until next time!