So What Smart Software Is Doing the Coding for Lagging Googlers?
January 13, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I read “Google Programmer Claims AI Solved a Problem That Took Human Coders a Year.” I assume that I am supposed to divine that I should fill in “to crack,” “to solve,” or “to develop”? Furthermore, I don’t know if the information in the write up is accurate or if it is a bit of fluff devised by an art history major who got a job with a PR firm supporting Google.
I like the way a Googler uses Anthropic to outperform Googlers (I think). Anyway, thanks, ChatGPT, good enough.
The company’s commitment to praise its AI technology is notable. Other AI firms toss out some baloney before their “leadership” has a meeting with angry investors. Google, on the other hand, pumps out toots and confetti with appalling regularity.
This particular write up states:
Paul [a person with inside knowledge about Google’s AI coding system] passed on secondhand knowledge from "a Principal Engineer at Google [that] Claude Code matched 1 year of team output in 1 hour."
Okay, that’s about as unsupported an assertion I have seen this morning. The write up continues:
San Francisco-based programmer Jaana Dogan chimed in, outing herself as the Google engineer cited by Paul. "We have been trying to build distributed agent orchestrators at Google since last year," she commented. "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."
So the “anonymous” programmer is Jaana Dogan. She did not use Opal, Google’s own smart software. Ms. Dogan used the coding tools from Anthropic? Is this what the cited passage is telling me?
Let’s think about these statements for a moment:
- Perhaps Google’s coders were doom scrolling, playing Foosball, or thinking about how they could land a huge salary at Meta now that AI staff are allegedly jump off the good ship Zuck Up? Therefore, smart software could indeed produce code that took the Googlers one year to produce. Googlers are not necessarily productive unless it is in the PR department or the legal department.
- Is Google’s own coding capability so lousy that Googlers armed with Opal and other Googley smart software could not complete a project with software Google is pitching as the greatest thing since Google landed a Nobel Prize?
- Is the Anthropic software that much better than Google’s or Microsoft’s smart coding system? My experience is that none of these systems are that different from one another. In fact, I am not sure that new releases are much better than the systems we have tested over the last 12 months.
The larger question is, “Why does Google have to promote its approach to AI so relentlessly?” Why is Google using another firm’s smart software and presenting its use in a confusing way?
My answer to both these questions is, “Google has a big time inferiority complex. It is as if the leadership of Google believes that grandma is standing behind them when they were 12 years old. When attention flags doing homework, grandma bats the family loser with her open palm. “Do better. Concentrate,” she snarls at the hapless student.
Thus, PR emanates PR that seems to be about its own capabilities and staff while promoting a smart coding tool from another firm. What’s clear is that the need for PR coverage outpaces common sense and planning. Google is trying hard to convince people that AI is the greatest thing since ping pong tables at the office.
Stephen E Arnold, January 13, 2025
Comments
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
