A meme of Nicolas Cage is making the rounds: “Firing 12k people rises the stock by 3%, one rushed AI presentation drops it by 8%.”

Quite a bit of criticism this week directed at Google for Bard and its handling of the release of OpenAI’s ChatGPT. There are several related articles: Ron Amadeo tells us that Google employees criticize CEO for “dumpster fire” response to ChatGPT, Joe Wituschek reports that Google’s own employees think Bard, its ChatGPT competitor, is being ‘rushed’, and Bradley Guichard asks Has Google Already Lost The ChatGPT War?. For good measure, Emma Roth tells us that Some Googlers reportedly aren’t happy about Bard’s ‘rushed’ announcement. and  Jennifer Elias notes  Google employees criticize CEO Sundar Pichai for ‘rushed, botched’ announcement of GPT competitor Bard, to give you a view of the topic from multiple angles.

Several users have taken to the company’s internal “Memegen” forum to voice their disappointment with the rushed, botched, and myopic announcement (those are the three adjectives which appear frequently in these reports), with the above-noted meme being one of the more notable results. Some have even gone so far as to criticize CEO Sundar Pichai for his leadership and short-term focus on the stock market.

Bard, an AI chatbot that uses transformer technology and is Google’s putative response to ChatGPT, has faced criticism from both employees and outsiders. The negative news caused Google’s stock to lose almost $100 billion of its market value in one day. Google announced Bard in an attempt leap ahead of Microsoft’s ChatGPT-powered Bing search engine. 

Alphabet relies heavily on Google Search, which makes up 69% of the company’s sales and even more of its operating income. This gives Google tremendous pricing power, and its search revenues have nearly doubled since 2018. AI’s importance to Google and its future in the search market is huge. As such, Google must fight to maintain its dominance in the market and keep up with new technologies and features like ChatGPT and Microsoft’s Bing.

Google has been developing new AI features for search for years, but has yet to push them to market, despite ChatGPT and other competitors gaining market attention. The irony here, of course, is that transformers, the technology underlying all current state-of-the-art large language models, was essentially invented at Google. Google seems to be turning to Xerox PARC and saying “hold my beer.”

Perhaps in response to all of this, Jennifer Elias writes that Google asks employees to rewrite Bard’s bad responses, says the A.I. ‘learns best by example.’ This is essentially an all-hands effort to provide the system with feedback to improve its performance. These quality testes are given these do’s and don’ts: 

Do’s

• Keep responses “polite, casual and approachable.”
• Use first person.
• Responses should have an “unopinionated, neutral tone.”

Don’ts

• Employees should “avoid making presumptions based on race, nationality, gender, age, religion, sexual orientation, political ideology, location, or similar categories.”
• Employees should not “describe Bard as a person, imply emotion, or claim to have human-like experiences.”
• Employees should “keep it safe,” by simply giving a “thumbs down” to answers that offer “legal, medical, financial advice” or are hateful and abusive, rather than trying to rewrite them.

The most exciting part is that those who participate will earn a “Moma badge.” This appears on the employee’s internal profile.