Microsoft CEO, Google Engineer Push Back On AI Quality Complaints

Microsoft Chief urges moving past “slop vs sophistication,” as Google engineer frames AI critics’ criticism as a sign of burnout in the tech industry.

Within one week of each other, Nadella and Dogan shared blog posts that shift the discussion away from the question of whether AI outputs are “good” or “bad” and to how people react towards the tech itself. 

Nadella: Move Beyond “Slop Vs Sophistication”

Nadella wrote in a private blog post, titled “Looking Ahead to 2026,” stated that the field should “get beyond the arguments of slop vs sophistication” and look at AI in the same way as “cognitive amplifier tools” that must now prove their the real-world benefits.

Days later, Dogan, a Principal Engineer working on Google’s Gemini API, posted on X that:

“people are only anti new tech when they are burned out from trying new tech.”

The timing is in line with Merriam-Webster declaring “slop” its Word of the Year. Publishers, the statements may not be as reassuring, but rather as a way to minimize the quality of their products.

Dogan: Criticism As “Burnout From Trying New Tech”

Jaana Dogan, a Principal Engineer working on Google’s Gemini API, added a more personal take in a widely shared post on X:

“People are only anti new tech when they are burned out from trying new tech. It’s understandable.”

People are only anti new tech when they are burned out from trying new tech. It’s understandable.

— Jaana Dogan ヤナ ドガン (@rakyll) January 4, 2026

Her comment frames a lot of AI skepticism as emotional fatigue rather than as a primarily technical or economic critique.​

The context is important. Just days earlier, Dogan described using Claude Code to generate a working prototype of a distributed agent orchestrator from a textual problem description in about an hour, closely mirroring patterns her team had spent roughly a year developing internally.

She noted that:

 “In 2023, I believed these current capabilities were still five years away.”​

I'm not joking and this isn't funny. We have been trying to build distributed agent orchestrators at Google since last year. There are various options, not everyone is aligned… I gave Claude Code a description of the problem, it generated what we built last year in an hour.

— Jaana Dogan ヤナ ドガン (@rakyll) January 2, 2026

Responses to her burnout post, however, highlighted a different side: users pointed to forced AI integrations, added subscription costs, privacy risks, and tools that feel unreliable or disruptive inside existing workflows.

Dogan is speaking in a personal capacity, but her role at Google makes the framing especially notable given how heavily Google is betting on Gemini across its product stack.​

The Quality Gap Publishers Are Dealing With

Google has long promoted E-E-A-T expectations–experience, expertise, authoritativeness, and trustworthiness–especially for “Your Money or Your Life” topics.

While AI products increasingly answer questions directly, with citations that can be hard to assess quickly.

What The Data Shows On AI Summaries And Clicks

The data on traffic suggests a significant impact. Pew Research Center found that when AI Overviews were displayed in Google results, just 8% of users clicked on a hyperlink, as opposed to 15% when there was no AI summary was shown, a 46.7% decrease. 

Similarweb data shows the proportion of search results related to news that don’t result in a click-through to news websites increasing from 56% to 69%. 

Additionally, Cloudflare estimates Google Search’s crawl-to-referral ratio to be around 14:1, with a ratio of 1700:1 in the case of OpenAI as well as 73,000 for Anthropic.

Publishers have for a long time accepted crawling for the sake of the ability to get traffic as well as visibility. Since AI capabilities answer more questions directly, many believe that the bargain is being eroded Content is still being used to train and fuel AI but a lower portion of people actually access the original websites.

Why The Messaging Shift Matters

Nadella’s and dogan’s posts provide hints as to the way AI quality debates could be addressed in 2026 less as a concern of accuracy, reliability and economics, but more as a matter of perceptions and fatigue among users. 

However, the fact that traffic is declining and the widening of crawl-to-referral gaps indicate that the economic implications are tangible.

Looking Forward

As time goes on, expect to see more communication that characterizes AI critiques as being a reaction issue instead of a problem with the ecosystem or a product. 

The question remains whether the major AI and search companies will alter the design of their products and distribution models to respond to increasing user and feedback from publishers.