00:00When you joined us at Invest, you said to me something to the effect that when we talk about
00:04innovation, very hard for the horizontal players to respond. OpenAI and Anthropic, for example,
00:10you said the models that have been trained on web training, they will take us very far,
00:14but not far enough. Do you think some of what we're seeing is kind of falls into your thinking?
00:20Although we did, we should point out that Amazon is now working with OpenAI. How do you see this
00:24story? Because I think we're trying to make sense of it. Thank you, Carol and Tim, and good to see
00:30you both. So what is transpiring is a first mover, in this case, OpenAI, experiencing something that
00:38we have not seen in tech in a long time, getting the followers to be truly fast followers and
00:43actually take market share. So it's very interesting to watch because OpenAI came right off the gates
00:49with chat GPT and had huge usage, largely driven by curiosity and interest and hype. And then you
00:58have Anthropic, which for a long time was viewed as a distant second, come right off the gates with
01:04the notions of trust, and they were going to be compliant and the use of their models and output
01:11was going to fit into enterprise workflows. And then the third player, Google, they have always been
01:17consumer oriented. So OpenAI went for it all. Anthropic focused on enterprise and Google focused
01:24on consumer. And Google, with a billion dollar consumer channel, had a real advantage. The reality
01:30has so far shows that those two players have caught on and are perhaps holding back the growth of OpenAI.
01:40Regina, one question that I think a lot of people have when it comes to this technology,
01:44at least from a consumer perspective, and look, this is where we sort of barely scratch the surface
01:47with what these models can do. But if you think about it from the perspective of consumers using
01:52these as enhanced search, maybe not paying, are they commodities because they can all sort of answer
01:57our questions in the same way? I think if you're only using chat GPT or cloud coworker, even the chat
02:07cloud is search engines, you are really missing the point from a consumer point of view. There is so
02:14much workflow automation that one can do. There are so many mundane tasks that you can actually have
02:20these platforms far beyond the what the weather will be like, or, you know, Google search, we have
02:27hopped a generation of capabilities. So go ahead. I don't think everybody gets that. I mean, I will tell
02:33you, I will admit, I tried to vibe code some elements of our program and to make our very heavy
02:40script more automated, where we do a lot of manual work. It doesn't, it didn't work for me. So, you
02:47know, there, there's a lot of people missing out. I completely agree. So I think a couple of
02:52suggestions. One, if you have a 20 year old or younger, have them for you. I talked to my 23
03:00year old
03:00the line. Maybe, maybe go ahead, go ahead. But more, but more fundamentally, Tim, first of all,
03:07it's not the chat function that will probably help you get better. At least in the context of
03:12cloud, it would be the cowork capability. So I will go, let me put my mom hat on. I have
03:18a theater
03:19oriented child. I always want to know when there are auditions, when there are plays. So I have built
03:25all of that on coworker purely on prompt and not a single line of code. And I've even asked it
03:30to
03:30generate the content in a tile format. So that is something that I enjoy, you know, ingesting and
03:37digesting as information and it's actionable, very, very basic and yet very powerful. So the number one
03:45sort of mindset should be what task can, you know, the open AI apps or the cloud, you know,
03:52various uses can actually generate for me or take over for me.
03:56So open AI did say, Rudina, just to kind of start to wrap up here, the company continues to see
04:02its
04:02push for more computing capacity as a great enabler, allowing it to deliver a better product
04:06experience to our customers. I mean, is open AI, should we be worried that maybe it's a lot of
04:12money chasing something that isn't going to pay off ROI, if you will, in the longer term, or we
04:19shouldn't go that far yet?
04:21The fundamental question is, what is the something? If we are talking about horizontal
04:26platforms, I think it's fair, at least in this moment in time, I think it's fairly clear that it
04:31will be winners take most, not a single winner, and that we will see specialization along consumer
04:38and enterprise and even within these markets, perhaps, you know, smaller verticals that I think
04:44will persist. When you look at the enterprise workflows, and particularly vertical industries,
04:50I think we will see a number of new players that we probably don't even know the names of,
04:54but I do, they've not made it to the public markets yet, that will really have specialized
04:59data, specialized architectures, and truly delivered high, high ROI productivity and beyond
05:04for the enterprise market.
05:06So, one last question. So, there might be a world, as we continue to see this build and spend
05:10and attention spent on AI, you know, you said to me about customer promiscuity, you know,
05:17customers use Gemini, next using ChatGPT Enterprise, and next day Claude, you know, we're using everything
05:24because it's no high cost to switching around, but that will not continue. So, there will be some
05:29fallout. These models, all of these large LLMs might not be around in just got about 40 seconds.
05:36I think they will be around. I don't know that they will be in the same form that we see
05:40them
05:40today. And then you will also see highly performing new models, new labs, and new players.
Comments