MongoDB CEO says if AI hype were the dotcom boom it is 1996
- Reference: 1726237994
- News link: https://www.theregister.co.uk/2024/09/13/mongodb_ceo_says_if_ai/
- Source link:
Speaking to investors at a Goldman Sachs event earlier this week, the boss of the NoSQL database company argued the tech industry was still early in the rise of AI, a development which - we're all told - will transform business and enterprise technology.
"I believe that AI is not a question of if, but when. I view the world we're in today [compared with] circa 1996, maybe 1997. Netscape was just launched a couple of years earlier. People were excited about the web, but the web was still very basic, static web pages, and it wasn't that interesting. People [were] starting businesses on the internet, maybe Amazon, eBay and a few others, but you weren't seeing a plethora of companies exploding onto the market," he said.
[1]
"I think in some ways, we're kind of at the same stage with the AI era," Ittycheria said.
[2]
[3]
The CEO's view on the trajectory of the AI boom is not in line with that of other observers.
In its Hype Cycle for Emerging Technologies, 2024 edition, released last month, analyst company Gartner [4]said GenAI is about to enter the dreaded "trough of disillusionment." Arun Chandrasekaran, Gartner distinguished VP analyst, told The Register at the time that expectations and hype around GenAI were "enormously high," but maintained it would have a "long term impact."
[5]
Meanwhile, ServiceNow chief financial officer [6]Gina Mastantuono admitted last week there may be a spending bubble surrounding GenAI.
Enterprises are [7]also struggling with the business case for GenAI projects more than a year after the boom started, global datacenter and colocation provider Equinix recently found.
MongoDB's Ittycheria told investors at the Goldman Sachs event there are three groups of use cases for GenAI at the moment: chatbots, research, and summarization and automation.
[8]
Quite where we are on the analogy with the dotcom boom and bust remains open to question. The CEO is correct in saying that as of 1996, Netscape Navigator – the browser which quickly dominated the nascent worldwide web – was a couple of years old, more or less. Amazon and eBay had also been formed in the previous two years.
[9]MongoDB takes a swing at PostgreSQL after claiming wins against rival
[10]AI boom is reshaping the face of cloud infrastructure
[11]Can't get Minecraft, MongoDB Cloud, others to work today? Blame that Azure outage
[12]Father of SQL says yes to NoSQL
Of course, the early dotcom era didn't end well. In early 2000, the Nasdaq Composite stock market index peaked after gaining 400 percent in the previous five years. It then went into a [13]thundering crash , [14]losing nearly two-thirds of its value in a year.
Whether we're heading for the same kind of crash, Ittycheria did not say. Still, he maintained people tend to "overestimate the impact of a new platform or technology in the short term but underestimate it in the long term."
Nonetheless, it could well be a bumpy ride, even if any dramatic changes for the worse are still five years off. In 2001, networking giant Cisco [15]wrote off $2.2 billion in inventory owing to stalled spending during the dotcom crash.
If you're making the picks and shovels for the current AI gold rush, the tough call is going to be figuring out when to stop. ®
Get our [16]Tech Resources
[1] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZuRhpOsilpP9azlCbOe0EgAAAUY&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZuRhpOsilpP9azlCbOe0EgAAAUY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZuRhpOsilpP9azlCbOe0EgAAAUY&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[4] https://www.theregister.com/2024/08/22/gartner_agi_hype_cycle/
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZuRhpOsilpP9azlCbOe0EgAAAUY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2024/09/04/servicenow_remains_bullish_on_gen_ai/
[7] https://www.theregister.com/2024/09/09/equinix_ai_business_case/
[8] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZuRhpOsilpP9azlCbOe0EgAAAUY&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[9] https://www.theregister.com/2024/08/30/mongodb_postgresql/
[10] https://www.theregister.com/2024/08/04/ai_cloud_infrastructure/
[11] https://www.theregister.com/2024/07/30/azure_outage_impact/
[12] https://www.theregister.com/2024/05/10/sql_cocreator_nosql/
[13] https://www.theregister.com/2000/04/17/dotcom_genocide_in_wall_st/
[14] https://www.theregister.com/2001/01/16/dotcom_value_drops/
[15] https://www.latimes.com/archives/la-xpm-2001-may-09-fi-61099-story.html
[16] https://whitepapers.theregister.com/
Currently, commercially speaking, worse than an FAQ section.
The point of chatbots is to save money on returns. If you buy something for £50 and there is something wrong with it, you go to customer service chat and the game is to get you to give up. To think that the effort is not worth getting £50 back. If you still getting through the hoops, the next step is to get you onto a minimum wage rep from the other side of the world, who typically don't even know where they work. So you have to explain everything again until you ask to talk to management or something. They need to put some road blocks there as well, but once you get there you typically get what you came for. The "management" is typically just a regular customer service person gate kept by chatbot etc.
The problem for research is that your LLMs make things up
Most people don't care. Even if you submit a paper, the reviewers unlikely are going to read it, but they will put it through automatic system of some sort that is also going to make mistakes.
jobs that LLMs are -or ever will be- capable of doing
At present I am certain LLM can replace very much any junior developer or any job that requires searching and "making sense" of information from large amount of text.
> MongoDB CEO says if AI hype were the dotcom boom it is 1996
The difference is that we don't actually have AI yet. We are not even close to having AI yet.
At least the dotcom boom actually had "the technology" in place before it attempted to monetise it.
Looking at the three use cases here:
chatbots
For fake social interactions chatbots are alright but as soon as they're on your company site giving your visitors actual information they're liable to either make things up or just return nothing, neither of which are useful. Currently, commercially speaking, worse than an FAQ section.
research and summarization
The problem for research is that your LLMs make things up, which means your research will be wrong unless you're actually doing research into LLMs. Meanwhile for summarisation a recent Australian study showed LLMs to be [1]worse then humans in every way for document summarisation, so...
automation
Machine Learning sits under the general headline of "AI" and might be useful for this. LLMs are only really good for automating interactions with people who you don't care about, as they're obviously LLMs and people will recognise that very quickly.
There's a couple of things I don't see here: jobs that LLMs are -or ever will be- capable of doing, and anything which will bring in an income within orders of magnitude of the levels of investment in AI we have seen recently.
Very hard to see it as anything other than a bubble at this point and I'd say that in his analogy we're closer to 1999.
[1] https://pivot-to-ai.com/2024/09/04/dont-use-ai-to-summarize-documents-its-worse-than-humans-in-every-way/