sorry if the last newsletter got caught off and I kept you hanging without Microsoft’s CEO’s quote 😂. I swear it wasn’t done on purpose to keep the suspense. Here’s what he said: A race starts today. We’re gonna move, and we’re gonna move fast.
I try to write Tech Pizza at the last possible moment so you get the freshest news, but sometimes that means that rush gets in the way. Forgive me, I’m just a human (sometimes I may get some help from ChatGPT though).
Anyway, this week’s news:
Let’s get started 🕺
Last week we talked about how Google lost $100B because of an error in their demo of Bard, their version of ChatGPT.
We also praised Microsoft because they integrated ChatGPT in both their search engine (Bing), and in their browser (Edge).
Now people found out that there were major errors in Bing’s demo too. Actually, much bigger problems than Google’s.
I’ll just show you the worst mistakes. Bing’s demo showed how it could summarize Gap’s Q3 2022 financial report. Bing answered in a realistic way but was often totally wrong. A few examples:
Bing: “Gap Inc. reported gross margin of 37.4%, adjusted for impairment charges related to Yeezy Gap, and merchandise margin declined 370 basis points.”
Truth: That’s the unadjusted gross margin. The gross margin adjusted for impairment charges was 38.7%. And the merchandise margin declined 480 basis points if we’re adjusting for impairment charges.
It gets worse.
Bing: “Gap Inc. reported operating margin of 5.9%, adjusted for impairment charges and restructuring costs, and diluted earnings per share of $0.42”
Truth: all these numbers are made up. The operating margin including impairment is 4.6% and excluding impairment is 3.9%. Adjusted diluted earnings per share is $0.71 and unadjusted is $0.77.
Bing: “Gap Inc. reaffirmed its full year fiscal 2022 guidance, expecting net sales growth in the low double digits”
Truth: Gap expects net sales to be down mid-single digits.
At some point, Bing is asked to compare Gap to Lululemon’s financial statements, and Bing compared made-up numbers to made-up numbers.
What do we learn from this?
The first thing is that the media’s narrative is controlling people’s sentiments around AI. According to the media (and therefore to most people), OpenAI is God and Google is struggling. The reality is that no one is God and everyone should struggle a bit more.
What’s absolute insanity is that this wasn’t a live demo: Microsoft had the time to check and fix the errors and they didn’t do it.
The second thing we learn is that we still have a long way to go. I still believe that this technology will change the world, but it will take time, especially for an industry like search where AI’s problems are particularly dangerous.
Sometimes it’s good to stop for a second, take a cold shower, and re-center expectations, otherwise, we risk getting another AI winter once we realize that this technology wasn’t God after all.
The sweet part of the newsletter: fun news from the crazy world of tech.
I totally recommend it, great info and also great fun! As you can deduct from above, I declare myself a fan🙂 happy to read this condensed newsletter about tech!
Cultural Change Director - E.ON
I feel more empowered intellectually 🙂
Business development - FNA SPA
Your 10 min shortcut to understand how tech impacts society and business, available in audio, video or text format
The attention around the release of GPT-4 is a million times more than I have seen with GPT-3. What it can do, why now, and what to expect.
Llama, a version of GPT-3 open-sourced by Meta. An algorithm like this has potentially destructive outcomes for society...