Goldman Sachs publishes blistering report on the AI bubble
July 10, 2024
Good morning. We’re looking at a bunch of accidentally related things in today’s edition.
Even as Goldman Sachs and others are beginning to pay more attention to the AI bubble that seems to be floating all around us, funding to AI startups last quarter spiked (by a lot) and OpenAI is forging ahead with a tech health company, that, in the context of an environment of high cost and virtually nonexistent revenue, feels like a bit of a risk.
We break it all down below.
— Ian Krietzberg, Editor-in-Chief, The Deep View
In today’s newsletter:
AI for Good: Accurately diagnosing ear infections
Source: Unsplash
Physician-scientists at the University of Pittsburgh recently designed a mobile app that leverages artificial intelligence to more accurately diagnose ear infections (known also as acute otitis media, or AOM).
The details: Researchers started by building a training library, which consisted of hundreds of videos of eardrums.
The tool is able to make a diagnosis after reviewing a short video, which is achieved by attaching an otoscope (one of those tools doctors use to look in your ears) to a smartphone camera.
The model was highly accurate, achieving a sensitivity and specificity value of more than 93%, suggesting that the tool is “more accurate than many clinicians.”
Why it matters: Senior author Dr. Alejandro Hoberman said that AOM — which can be difficult to “discern from other ear conditions without extensive training” — is often misdiagnosed.
“Underdiagnosis results in inadequate care and overdiagnosis results in unnecessary antibiotic treatment, which can compromise the effectiveness of currently available antibiotics. Our tool helps get the correct diagnosis and guide the right treatment,” he said.
Prepare for high-paying AI-enabled roles
Ready to transition to top AI/ML roles or get AI-enabled in your domain?
Interview Kickstart's courses help you build practical AI/ML skills taught live by FAANG+ experts and industry leaders.
✅Hands-on capstone projects and case studies
✅Taught live by FAANG+ AI expert instructors
✅Tailored for engineers and managers
✅1:1 Career support
OpenAI teams up with Arianna Huffington to launch AI health startup
Source: Unsplash
OpenAI CEO Sam Altman and Huffington Post founder Arianna Huffington are teaming up to bring AI to healthcare in the shape of a new venture: Thrive AI Health.
The details: The new company — which will be funded by OpenAI’s Startup Fund and Thrive Global — aims to create a hyper-personalized health coach powered by AI.
The mission, according to a Time op-ed co-authored by Altman and Huffington, is to use AI to drive “behavior change” across a few key areas, including sleep, food and exercise.
The model will be trained on “the best peer-reviewed science” in addition to any personal information users feel inclined to share.
The model will, according to Altman and Huffington, give “precise” health recommendations, such as: “swap your third afternoon soda with water and lemon; go on a 10-minute walk with your child after you pick them up from school at 3:15 p.m.; start your wind-down routine at 10 p.m. since you have to get up at 6 a.m. the next morning to make your flight.”
The two argue that this will solve the unequal distribution of healthcare access; while some, they argue, have access to life coaches and trainers, now the masses can use AI to “scale and democratize the life-saving benefits of improving daily habits and address growing health inequities.”
My view: I don’t like the term ‘democratize access.’ You hear it a lot in this field, and I find it exceptionally misleading. It’s not clear what Thrive AI Health will cost, but you can be sure it won’t be free. Plus, there will be regular, and likely large, carbon emissions associated with it.
Where this seems to land you — based on the examples Altman and Huffington provided — is a more costly version of the internet; I can do a quick (free) search right now to find hundreds of tips on ways to live healthier. Anyone who does that should know to drink water rather than soda, and to walk more. This information is out there (books!). It’s not clear what (novel, accurate, reliable) service this would accomplish, and who it would democratize access to.
This wouldn’t help the nearly three billion people around the world who don’t have internet access, for instance.
In the context of our main story today, this is, well, let’s just call it intriguing.
Share 2 Days with the Brightest Minds in AI including scientists and engineers from OpenAI, Meta, Deepmind, and many more.
This week only — Register using discount code: deepview24 for a $350 discount!*
U.S. Nuke Agency Buys Internet Backbone Data (404 Media).
Andreessen Horowitz is building a stash of more than 20,000 GPUs to win AI deals (The Information).
China leads the world in adoption of generative AI, survey shows (Reuters).
Etsy CEO says company is escaping ‘race to the bottom’ and getting back to its artisan roots (CNBC).
Google is no longer claiming to be carbon neutral (Bloomberg).
Funding to AI startups ‘more than doubled’ in Q2
Source: Crunchbase
Startups around the world experienced a bit of a surge in funding in the second quarter of 2024, according to Crunchbase data published Tuesday.
The details: Crunchbase found that total funding for the quarter reached $79 billion, a 16% increase from the previous quarter.
AI represented the largest sector in the quarter; funding to AI companies — at $24 billion — made up some 30% of the total.
This spike in AI funding marks the largest single-quarter raise in recent years. It came in at more than double the funding the sector received the previous quarter.
Why it matters: The surge in funding comes — as Crunchbase noted — despite an “uncertain” market.
Concerns about revenue generation have been mounting, as we break down below. Analysts at Citigroup, meanwhile, said recently that it’s time for investors to start taking profits in the biggest AI names, expecting “significantly more volatility” ahead.
But the venture side of things seems as excited about AI as ever. It will be interesting to watch the way funding data swings over the next quarter or two.
Effortlessly Schedule Developer Interviews with PestoAI
Imagine effortlessly adding top-tier developers to your interview calendar.
PestoAI can source and interview developers from around the globe, outperforming the average recruiter. You only pay when you succeed.
Using PestoAI feels like magic. Here's what happens behind the scenes:
Targeted Sourcing: PestoAI curates candidates based on your job description, using a model trained by local market recruiters.
Company Pitch: Pesto pitches your company to the best-fit developers.
Technical Screening: Candidates undergo a thorough technical interview using our proprietary technology.
Seamless Scheduling: PestoAI schedules the interviews directly on your calendar.
For Deep View readers, PestoAI is offering an exclusive 50% OFF if you sign up today!
Goldman Sachs publishes blistering report on the AI bubble
Source: Created with AI by The Deep View
In a bit of a continuation of our piece this week about the math behind the AI bubble — and the roughly $600 billion hole that the industry finds itself in today — is a recent report released by Goldman Sachs titled “GenAI: Too much spend, too little benefit?”
The report, as Ed Zitron so rightly pointed out, is significant in that “Goldman Sachs, like any investment bank, does not care about anyone's feelings unless doing so is profitable.” The company’s turn — from suggesting in May that AI was "showing very positive signs of eventually boosting GDP and productivity” — to now, indicates that Goldman is seeing the same signs that Sequoia and everyone else are seeing.
And they’re not sold on the hype.
The details: The 30-page report begins with a simple sentiment: “Tech giants and beyond are set to spend over $1 trillion on AI capex in coming years, with so far little to show for it.” The report is then broken up into a series of interviews with Daron Acemoglu, an economist at MIT, and Jim Covello, Goldman’s Head of Global Equity Research, in addition to a few others.
Acemoglu — who predicted in a recent paper that, over the next decade, AI will lead to a 0.5% increase in productivity and a 1% increase in GDP — believes that while generative AI has the potential to transform the process of scientific discovery, he doesn’t expect these kinds of changes to appear within the next 10 years.
He also questioned the idea of an eventual superintelligence, saying that the primary use of current AI involves targeted automation, which isn’t great if such tech is used prematurely.
“Too much automation too soon could create bottlenecks and other problems for firms that no longer have the flexibility and trouble-shooting capabilities that human capital provides,” he said.
Acemoglu additionally challenged the idea of “scale is all you need” that has been touted by many in the industry, going further than just suggesting that scale is *not all you need to suggest that the framing fundamentally makes no sense.
“What does it mean to double AI’s capabilities? For open-ended tasks like customer service or understanding and summarizing text, no clear metric exists to demonstrate that the output is twice as good,” he said.
“Similarly, what does a doubling of data really mean, and what can it achieve? Including twice as much data from Reddit into the next version of GPT may improve its ability to predict the next word when engaging in an informal conversation, but it won't necessarily improve a customer service representative’s ability to help a customer troubleshoot problems with their video service.”
He said that a “big leap of faith is still required to believe that the architecture of predicting the next word in a sentence will achieve capabilities as smart as HAL 9000 in 2001: A Space Odyssey. It’s all but certain that current AI models won’t achieve anything close to such a feat within the next ten years.”
Covello, meanwhile, summed up his skepticism with a simple question: Considering the $1 trillion(+) that the industry is expected to spend on AI infrastructure over the next few years, “What $1 trillion problem will AI solve?”
“Replacing low-wage jobs with tremendously costly technology is basically the polar opposite of the prior technology transitions I’ve witnessed in my thirty years of closely following the tech industry.”
Covello said that AI is fundamentally unlike the early days of the internet, since the internet was a low-cost solution that replaced costly (brick-and-mortar) incumbent solutions, where AI is a super high-cost ‘solution’ to a low-cost system.
Covello isn’t just skeptical about the ROI offered by AI; he’s skeptical that it’ll actually be as transformative as enthusiasts have been claiming. The internet, he said, was transformative — current AI, at best, is just adding a layer of efficiency to existing processes, at a much higher cost.
“For example, we’ve found that AI can update historical data in our company models more quickly than doing so manually, but at six times the cost,” Covello said.
“This is not a matter of just some tweaks being required here and there; despite its expensive price tag, the technology is nowhere near where it needs to be in order to be useful for even such basic tasks,” he said. “And I struggle to believe that the technology will ever achieve the cognitive reasoning required to substantially augment or replace human interactions.”
I highly recommend reading Covello’s full thoughts (on pages 10 and 11) but will include one more snippet from his interview:
Where the transformative potential of the internet and smartphones was understood right at the very beginning, Covello said that “no comparable roadmap exists today.”
“AI bulls seem to just trust that use cases will proliferate as the technology evolves. But eighteen months after the introduction of generative AI to the world, not one truly transformative — let alone cost-effective — application has been found.”
AI researcher Grady Booch added in response to the report that “GenAI also represents a tremendous opportunity cost. Imagine if all the smart folk, all the hardware, all the investments were not focused on this dead-end architecture.”
It seems likely that what we are experiencing right now is a bubble. The costs are too high, and the use cases aren’t quite there yet, a combination that leads to an unsustainable environment.
But a bubble doesn’t mean that AI is useless — some of the companies that survived the dot-com bubble (Amazon, for instance) have spent the past 20 years becoming some of the largest, most consequential companies in the world.
The interesting challenge of today’s environment is figuring out which companies — and what kinds of companies/use cases — will survive the bubble.
I wonder how much this industry will spend, in both dollars and carbon emissions, before it understands that believing something doesn’t make it real.
Which image is real? |
A poll before you go
Thanks for reading today’s edition of The Deep View!
We’ll see you in the next one.
Your view on AI in healthcare:
Of the clinicians who voted, half said there would be good and bad, while a third said it would be terrible.
Of the non-clinicians, more than half said there would be good and bad, while a third said it sounds like a good idea. Only 14% of you (non-clinicians) would have a problem with your doctor using AI.
Not a clinician, but this sounds like a good idea:
“I've thought for a few years that it would be great to have AI receiving the same information the docs receive for patients, with all medical history etc and basically being the backup - from time to time believe it might catch something that docs (who in fairness are more hurried across more disciplines than before) could overlook.”
Would you pay for access to a personalized AI health coach? |