Once AI is more reliable, I think this is likely to turn out to be true. It would be much easier than the way we currently search for information, including our online shopping searches.
Bill Gates, Microsoft Corp’s co-founder, on Monday said the technology race to win is the development of the top artificially intelligent agent, poised to disrupt search-engine, productivity and online shopping sites.
“You’ll never go to a search site again,” he said. “You’ll never go to Amazon.”
Speaking at AI Forward 2023, an event in San Francisco hosted by Goldman Sachs Group Inc and SV Angel, Gates said it would disappoint him if Microsoft were not in the running, though there was a 50% chance the top player to emerge will be a startup.
Google famously declared a “code red” situation when OpenAI launched GPT3. This totally makes sense, because for most people Google has been playing the part of being the gatekeeper for the internet. Google this, Google that, they have been the middleman and this technology cuts that role: if something is more useful that the search engine, people will flock to it overnight, the cost of switching is non existing. Microsoft has been swift by securing agreements with OpenAI and executing the integration with Bing in months? weeks?
It will be interesting to see how all this shakes out once we are past this initial free for all. Some that follow Microsoft are speculating on how much an AI enhanced MS Office subscription will cost. Some guesstimates range from $10/month to $20/month above the current cost of MS365.
It’s too early to see how this will affect search engines.
I think it’s far too soon to discount his take on this based on some cherry-picked things he didn’t get right in the past. He’s also gotten a lot right and is largely responsible for computing today as we know it (either directly or by spurring competition).
I’m no expert so in my mind he is just as likely to be right as he is wrong. What do I know? But I do think the takes out there that talk about AI in the form of productivity and personal assistants are really onto something. We just saw Spark release their AI powered proofreader (time will tell if it’s much different than Grammarly).
A personal assistant that I can use to schedule events, find time in my calendar for hanging out with friends, do my shopping for me, proofread my emails at work, etc.? Sign me up. I don’t think AI can replace humans in these regards completely, but my current position has me doing a lot of scheduling and prepping a VIP for their work…rough WAG: properly-implemented AI could replace half of what I do. Does that make me scared for my job? Nope (I’m government, also, so let’s be honest…it’ll take us a while to get there). But what would be cool is I could take that 50% of my time and point it in the direction of thinking big thoughts, strategic choice, etc.
I think history may show that Satya Nadella had the best crystal ball of any Microsoft CEO. Microsoft stock rarely topped $40 a share until Nadella took over in 2014. Today it closed at $315.26.
I’ve wanted one of those even since Siri was released as an app and we may see one in the next couple of years. But I don’t think the best of them will be running on device, at least not at first. They may need more horsepower than that available on current mobile devices.
Once Microsoft’s Co-Pilot for Office, and Google’s Duet AI for Google Workspace are available for testing we may have a better idea what to expect.
I think you are right, but less so because of stock price (in my very humble opinion (as someone who has money in the market) is primarily about growth and drives companies to do crazy things) and more about I like where he has taken the company in software and products (those two things can definitely be synonymous).
Completely agree. I need to finish the Upgrade episode and maybe re-listen to the new article that kicked it off, but I thought I heard the guys talking about companies running local/“air-gapped” instances of AI (I thought ChatGPT). There’s no way the government (in my case) or a company releases classified/proprietary information to a 3rd party. But a contracted piece of software running on approved in-house/contracted servers? Yup.
Like I said, I’m absolutely ready for the AI revolution in productivity/assistant software. Maybe I’ll come to eat those words in the future, but that is where my head is at right now.
Right! Given the computational resources needed to run inference on a LLM, it’s going to be very difficult to monetize a GPT-like engine based on ad “impressions” (whatever they may look like in a textual conversation). Perhaps a combination of very highly targeted ads with a subscription model for more powerful LLMs? But is it possible for a hypothetical Google killer to infer purchase motivations during a textual conversation with a bot? And better yet, Is it possible to do it significantly better than during a regular Google search session?
Not clear to me, but I’d say Google is working furiously on that.
