Google and Its PR Response to the ChatGPT Buzz Noise
January 16, 2023
A crazy wave is sweeping through the technology datasphere. ChatGPT, OpenAI, Microsoft, Silicon Valley pundits, and educators are shaken, not stirred, into the next big thing. But where is the Google in this cyclone bomb of smart software? The craze is not for a list of documents matching a user’s query. People like students and spammers are eager for tools that can write, talk, draw pictures, and code. Yes, code more good enough software, by golly.
In this torrential outpouring of commentary, free demonstrations, and venture capitalists’ excitement, I want to ask a simple question: Where’s the Google? Well, to the Google haters, the GOOG is in panic mode. RED ALERT, RED ALERT.
From my point of view, the Google has been busy busy being Google. Its head of search Prabhakar Raghavan is in the spotlight because some believe he has missed the Google bus taking him to the future of search. The idea is that Googzilla has been napping before heading to Vegas to follow the NCAA basketball tournament in incorrect. Google has been busy, just not in a podcast, talking heads, pundit tweeting way.
Let’s look at two examples of what Google has been up to since ChatGPT became the next big thing in a rather dismal economic environment.
The first is the appearance of a articles about the forward forward method for training smart software. You can read a reasonably good explanation in “What Is the “Forward-Forward” Algorithm, Geoffrey Hinton’s New AI Technique?” The idea is that some of the old-school approaches won’t work in today go-go world. Google, of course, has solved this problem. Did the forward forward thing catch the attention of the luminaries excited about ChatGPT? No. Why? Google is not too good at marketing in my opinion. ChatGPT is destined to be a mere footnote. Yep, a footnote, probably one in multiple Google papers like Understanding Diffusion Models: A Unified Perspective (August 2022). (Trust me. There are quite a few of these papers with comments about the flaws of ChatGPT-type software in the “closings” or “conclusions” to these Google papers.)
The second is the presentation of information about Google’s higher purpose. A good example of this is the recent interview with a Googler involved in the mysterious game-playing, protein-folding outfit called DeepMind. “DeepMind’s CEO Helped Take AI Mainstream. Now He’s Urging Caution” does a good job of hitting the themes expressed in technical papers, YouTube video interviews, and breezy presentations at smart software conferences. This is a follow on to Google’s talking with an MIT researcher Lex Fridman about the Google engineer who thought the DeepMind system was a person and a two hour chat with the boss of DeepMind. The CEO video is at this link.
I want to highlight three points from this interview/article.
[A] Let’s look at this passage from the Time Magazine interview with the CEO of DeepMind:
Today’s AI is narrow, brittle, and often not very intelligent at all. But AGI, Hassabis believes, will be an “epoch-defining” technology—like the harnessing of electricity—that will change the very fabric of human life. If he’s right, it could earn him a place in history that would relegate the namesakes of his meeting rooms to mere footnotes.
I interpret this to mean that Google has better, faster, cheaper, and smarter NLP technology. Notice the idea of putting competitors in “mere footnotes.” This is an academic, semi-polite way to say, “Loser.”
[B] DeepMind alleged became a unit of Alphabet Google for this reason:
Google was “very happy to accept” DeepMind’s ethical red lines “as part of the acquisition.”
Forget the money. Think “ethical red lines.” Okay, that’s an interesting concept for a company which is in the data hoovering business, sells advertising, has a bureaucratic approach I heard described as described as slime mold, and is being sued for assorted allegations of monopolistic behavior in several countries.
[C] The Time Magazine article includes this statement:
Back at DeepMind’s spiral staircase, an employee explains that the DNA sculpture is designed to rotate, but today the motor is broken. Closer inspection shows some of the rungs of the helix are askew.
Interesting choice of words: “The motor is broken” and “askew.” Is this irony or just the way it is when engineering has to be good enough and advertising powers the buzzing nervous system of the company?
From my point of view, Google has been responding to ChatGPT with academic reminders that the online advertising outfit has a better mousetrap. My thought is that Google knew ChatGPT would be a big deal. That realization sparked the attempt by Google to answer questions with cards and weird little factoids related to the user’s query. The real beef or “wood behind” the program is the catchy forward forward campaign. How is that working out? I don’t have a Google T shirt that spells out Forward Forward. Have you seen one? My research suggests that Google wants to corner the market on low cost training data. Think Snorkel. Google pushes synthetic data because it is not real and, therefore, cannot be dragged into court over improper use of Web-accessible content. Google, I believe, wants to become the Trader Joe’s of off-the-shelf training data and ready-to-run smart software models. The idea has been implemented to some degree at Amazon’s AWS as I recall.
Furthermore, Google’s idea of a PR blitz is talking with an MIT researcher Lex Fridman. Mr. Fridman interviewed the the Google engineer (now a Xoogler) who thought the DeepMind system was a person and sort of alive. Mr. Fridman also spoke with the boss of DeepMind about smart software. (The video is at this link.) The themes are familiar: Great software, more behind the curtains, and doing good with Go and proteins.
Google faces several challenges with its PR effort to respond to ChatGPT:
- I am of the opinion that most people, even those involved in smart software, are not aware that Google has been running a PR and marketing campaign to make clear superiority of its system and method. No mere footnote for the Google. We do proteins. We snorkel. We forward forward. The problem is that ChatGPT is everywhere, and people like high school students are talking about it. Even artists are aware of smart software and instant image generation OpenAI style.
- Google remains ill equipped to respond to ChatGPT’s sudden thunder showers and wind storms of social buzz. Not even Google’s rise to fame matches what has happened to OpenAI and ChatGPT in the last few months. There are rumors that Microsoft will do more than provided Azure computing resources for ChatGPT. Microsoft may dump hard cash billions into OpenAI. Who is not excited to punch a button and have Microsoft Word write that report for you? I think high school students will embrace the idea; teachers and article writers at CNet, not so much.
- Retooling Google’s decades old systems and methods for the snappy ChatGPT approach will take time and money. Google has the money, but in the world of bomb cyclones the company may not have time. Technology fortunes can vaporize quickly like the value of a used Tesla on Cars and Bids.
Net net: Google, believe it or not, has been in its own Googley way trying to respond to its ChatGPT moment. What the company has been doing is interesting. However, unlike some of Google’s technical processes, the online information access world is able to change. Can Google? Will high school students and search engine optimization spam writers care? What about “axis of evil” outfits and their propaganda agencies? What about users who do not know when a machine-generated output is dead wrong? Google may not face an existential crisis, but the company definitely knows something is shaking the once-solid foundations of the buildings on Shoreline Drive.
Stephen E Arnold, January 16, 2023
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
