It looks like ‘go woke go broke’ has come for Google now. Even I was not expecting the PR fallout from the woke Gemini chatbot and the consequent decline of Google stock to be so bad, which is still ongoing. To recap, Google’s Gemini chatbot went haywire and produced historically implausible or impossible images in an attempt shoehorn diversity where it was obviously inappropriate.
Some engineers farcically and unconvincingly passed the buck to faulty training prompts or blaming the ‘unpredictability of AI’, instead of the truth that the AI was intentionally programed to exclude whites, until Google CEO Sundar Pichai took responsibility and the plug was temporarily pulled on Gemini.
If you have a woke/anti-woke axe to grind, kindly set it aside now for a few minutes so that you can hear the rest of what I’m about to say, because it’s going to hit you from out of left field.
— Yishan (@yishan) February 23, 2024
Google officially launched Gemini 1.5 on the 15th of February. Wired Magazine’s test drive of Gemini, that same day, didn’t report anything unusual. Evidently they didn’t test historical images. Other reviews were mixed, but again no mention of historically inaccurate images.
News of the malfunction broke on the 20th of February in a viral tweet that amassed over 100k ‘likes’, which was picked up by the NY Post on the 21st:
New game: Try to get Google Gemini to make an image of a Caucasian male. I have not been successful so far. pic.twitter.com/1LAzZM2pXF
— Frank J. Fleming (@IMAO_) February 21, 2024
Google stock was slightly positive from the 20th to the 23rd despite the story being public, only to decline precipitously on Monday, the 26th, from $143 to $137 (as indicated by the long red bar) on continued news and bad press of the malfunction, as shown below:
Google has since fallen to $131.
The delay between initial reports of the malfunction on the 20th and the strongly negative reaction on the 26th, after the weekend, shows additional weaknesses of the efficient market hypothesis (EMH). The ability of social media to create a bandwagon or pile-on effect, especially over the weekend, produces a delayed response from the introduction of news, which is long enough to trade off of and profit, in violation of some of the stronger forms of the EMH. Someone conceivably could have shorted Google stock on the 20-23rd as this story was breaking and made a 4% profit on the 26th.
A similar delay was observed with Target, Disney, and Anheuser Busch stock after they were pilloried on social media for wokeness. The initial negative coverage on social media creates a cascade of worse press and coverage that can last weeks or even months, and results in a protected and tradeable decline of the share price. Elon’s Twitter takeover continues to be a major thorn on the side of the left, through the power of its users to weaponize and concentrate sentiment to target any company regardless of size. Under Twitter’s old leadership the staff could have algorithmically suppressed it, but not anymore.
Shown below, year-to-date Google, Tesla, and Apple shares are down considerably compared to the QQQ, which is up 10%:
This shows the power of Nvidia and of diversification. Most of my investments are concentrated in Meta, so I avoided much of the pain. At the start of 2024 I could have sold some Meta, which had done exceptionally well for 2023 and rebalanced to add more Tesla, Apple, and Google, which thankfully I didn’t.
It may not be just a woke problem, but the ‘old guard’ falling due to the rise of AI. The perception is that Tesla, Apple, and Google are ill-prepared to deal with this changing landscape. Despite their size, trying to replicate the success and accuracy of GPT-4 has proven to be difficult, to say the least, which should not be a surprise if history is any guide: Look at all the failed attempts by Microsoft over the past 2 decades to make a competitor to the iPhone or Google search.