ChatGPT Seen Launching an AI App Explosion
Discover more from AI in Business
ChatGPT Seen Launching an AI App Explosion
Analysis of ChatGPT’s pricing shows its work is available at the rate of 13 cents/hour; Microsoft execs endorse the idea of government regulation of AI
By John P. Desmond, Editor, AI in Business
The release of the ChatGPT API by OpenAI on March 1 sets up the industry for an explosion of AI-based applications, the likes of which literally have never been seen before.
To answer the question of how smart is ChatGPT, attempts have been made to quantify its IQ. In a recent account on Reddit from TotalPositivity, the author estimated ChatGPT’s IQ as between 83 and 147, quite a range.
“To me, it is essentially a very good “apprentice” level of intelligence. I wouldn’t let it rewire my house or remove my kidney, yet it would be better than me personally at advising on those things in a pinch where a professional is not available,” TotalPositivity stated. .
The OpenAI announced pricing for ChatGPT was set at about 500,000 tokens per dollar, according to TotalPositivity. If a human thinks at the rate of 800 words per minute, and a word is about 1.33 tokens, it means a human working 40-hour weeks for a year could produce 132 million tokens per year of thought. This would cost $264 from ChatGPT, according to TotalPositivity’s calculations.
Further, if the global workforce of 3.32 billion people could produce 440 quadrillion tokens per year, it could cost about $882 billion. Said another way, “You can now purchase an intellectual workforce the size of the entire planetary economy, maximally employed and focused, for less than the US military spends per year,” Total Positivity calculated.
And said another way, it is as if a nation the size of India and China combined dropped into the Pacific full of workers, who all work remotely, always pay attention, and cost only $264/year or 13 cents per hour. “Whatever future you’ve been envisioning, today may forever be the anniversary of all of it.,” the author concluded.
Brad Smith on 60 Minutes Endorses Government Regulation of AI
Meanwhile, Microsoft President Brad Smith, in an interview broadcast on 60 Minutes on March 5, discussed the interview published in The New York Times with the Bing chatbot powered by ChatGPT that went awry and was roundly criticized. He stated, “I think it would be a mistake if we were to fail to acknowledge that we are dealing with something that is fundamentally new. This is the edge of the envelope.”
Within 24 hours, Microsoft made a fix that would prevent a two-hour conversation with Bing; a new guardrail was constructed.
When asked by Leslie Stahl what could be done to prevent a company from releasing a ChatGPT-based product that could be seriously misused, Smith stated, “We’re going to need laws … I think it’s inevitable.”
Stahl asked if Microsoft would support a federal agency to regulate the technology industry, similar to the FAA for airlines and the FDA for pharmaceuticals. Smith stated, “I think I probably would. I think that something like a digital regulatory commission, if designed the right way, could be precisely what the public will want and need.”
(Ed. Note: It seems weird that a tech company would release a product to the world, that it then asks the government to regulate, to keep it from going out of control. One has to wonder if this is asking too much of a Congress that cannot come together on so many basic issues, like voting rights.)
Thousands of Developers Working on AI at Microsoft
A similar theme was struck by Eric Boyd, corporate VP of Microsoft’s AI platform, in a recent account published in MIT Sloan Management Review. Commenting on the role of government in relation to AI, Boyd stated there is “going to be a place where there has to be some government regulation of when it comes to what we should or shouldn’t do as societies. If you leave that up to corporations, corporations will make different decisions. That has to become a government decision.”
Meanwhile, AI development continues at a blistering pace inside Microsoft. Boyd ticked off some examples: “The speech quality that we deliver through our speech API literally improves every month … Our vision models have just exploded in quality recently … The large language models are just incredibly powerful these days, and so [there’s] just a massive explosion going on there.”
On the impact of ChatGPT: “We’re starting to see all the applications that we’re going to be able to light up as a result of that … It’s going to change everything. It’s going to change all the ways that we interact with computers, and so that’s really exciting to see.”
Thousands of developers and researchers across Bing and Office are training models “to do things that you’ll experience every day as a user of Microsoft,” Boyd stated. Much of this AI will be embedded within products, not highly visible. “I want us to think about scenarios and products where the portion of people using it, who were using the AI-powered features, is 100 percent,” Boyd stated.
It may not be possible to avoid AI, because it is “intrinsic to the product,” Boyd stated, noting, “Search is, of course, like that. You can’t avoid using AI in search. When you are talking to your phone to compose a text message with speech, you’re using AI 100 percent of the time you do that. And so increasingly, as we see these scenarios, there are so many things that are just not possible [without AI],” he stated.
Finally, “There’s this whole field of AI-powered applications that is really about to start to blossom, where the application just doesn’t exist without the AI,” Boyd stated.
Citing an example of customer experience, Boyd mentioned CarMax, which is using ChatGPT-based models to summarize reviews on each car make and model, and then generating a report, a process that might have previously been impractical. “Now they have this high-quality, valuable content that’s directing people to their site,” Boyd stated.
Like many companies, Microsoft is searching for the business model around ChatGPT. Advertising will be a major driver, but the company expects fewer ads than traditional search such as that offered by Google allows, according to Yusuf Mehdi, the head of Bing at Microsoft, in a recent account in The New York Times. “We’re going to learn that as we go,” Mehdi stated about how the new chatbot business model will generate revenue.
Meanwhile, Microsoft has started offering a cloud service built in conjunction with OpenAI that provides developers with coding suggestions, for $10 a month.
Google Also Probing for Chatbot Business Model
Google has also been exploring how to make money from a chatbot business model, with its recent introduction of the Bard chatbot. Dan Taylor, a Google VP of global advertising, stated that he sees the technology as “experimental” and is focused on using a large language model to improve traditional search.
“The discourse on AI is rather narrow and focused on text and the chat experience,” Taylor stated. “Our vision for search is about understanding information and all its forms: language, images, video, navigating the real world.”
Some entrepreneurs are envisioning a new era of online retail that upends the advertising-based model Google has built.
Manish Chandra, the chief executive of Poshmark, an online secondhand store, imagines a chatbot building profiles of an individual’s tastes, then recommending and possibly buying relevant clothes, electronics or groceries to make a recipe.
“It becomes your mini-Amazon,” stated Chandra to the Times. He has made integrating generative AI into Poshmark a top company priority over the next three years. “That layer is going to be very powerful and disruptive and start almost a new layer of retail,” he stated.
Read the source articles and information on Reddit, on 60 Minutes, in MIT Sloan Management Review and in The New York Times.
(Write to the editor here.)