The AI Revolution Is Already Losing Steam

Table of Contents

The AI Revolution Is Already Losing Steam

The AI Revolution Is Already Losing Steam

https://www.wsj.com/tech/ai/the-ai-revolution-is-already-losing-steam-a93478b1

By Christopher Mims May 31, 2024 9:00 pm ET

Nvidia reported eye-popping revenue last week. Elon Musk just said human-level artificial intelligence is coming next year. Big tech can’t seem to buy enough AI-powering chips. It sure seems like the AI hype train is just leaving the station, and we should all hop aboard. But significant disappointment may be on the horizon, both in terms of what AI can do, and the returns it will generate for investors. The rate of improvement for AIs is slowing, and there appear to be fewer applications than originally imagined for even the most capable of them. It is wildly expensive to build and run AI. New, competing AI models are popping up constantly, but it takes a long time for them to have a meaningful impact on how most people actually work. These factors raise questions about whether AI could become commoditized, about its potential to produce revenue and especially profits, and whether a new economy is actually being born. They also suggest that spending on AI is probably getting ahead of itself in a way we last saw during the fiber-optic boom of the late 1990s—a boom that led to some of the biggest crashes of the first dot-com bubble.

The pace of improvement in AIs is slowing

Most of the measurable and qualitative improvements in today’s large language model AIs like OpenAI’s ChatGPT and Google’s Gemini—including their talents for writing and analysis—come down to shoving ever more data into them. These models work by digesting huge volumes of text, and it’s undeniable that up to now, simply adding more has led to better capabilities. But a major barrier to continuing down this path is that companies have already trained their AIs on more or less the entire internet, and are running out of additional data to hoover up. There aren’t 10 more internets’ worth of human-generated content for today’s AIs to inhale.

To train next generation AIs, engineers are turning to “synthetic data,” which is data generated by other AIs. That approach didn’t work to create better self-driving technology for vehicles, and there is plenty of evidence it will be no better for large language models, says Gary Marcus, a cognitive scientist who sold an AI startup to Uber in 2016. AIs like ChatGPT rapidly got better in their early days, but what we’ve seen in the past 14-and-a-half months are only incremental gains, says Marcus. “The truth is, the core capabilities of these systems have either reached a plateau, or at least have slowed down in their improvement,” he adds. Further evidence of the slowdown in improvement of AIs can be found in research showing that the gaps between the performance of various AI models are closing. All of the best proprietary AI models are converging on about the same scores on tests of their abilities, and even free, open-source models, like those from Meta and Mistral, are catching up.

AI could become a commodity

A mature technology is one where everyone knows how to build it. Absent profound breakthroughs—which become exceedingly rare—no one has an edge in performance. At the same time, companies look for efficiencies, and whoever is winning shifts from who is in the lead to who can cut costs to the bone. The last major technology this happened with was electric vehicles, and now it appears to be happening to AI. The commoditization of AI is one reason that Anshu Sharma, chief executive of data and AI-privacy startup Skyflow, and a former vice president at business-software giant Salesforce, thinks that the future for AI startups—like OpenAI and Anthropic—could be dim. While he’s optimistic that big companies like Microsoft and Google will be able to entice enough users to make their AI investments worthwhile, doing so will require spending vast amounts of money over a long period of time, leaving even the best-funded AI startups—with their comparatively paltry warchests—unable to compete. This is happening already. Some AI startups have already run into turmoil, including Inflection AI—its co-founder and other employees decamped for Microsoft in March. The CEO of Stability AI, which built the popular image-generation AI tool Stable Diffusion, left abruptly in March. Many other AI startups, even well-funded ones, are apparently in talks to sell themselves.

Today’s AI’s remain ruinously expensive to run

An oft-cited figure in arguments that we’re in an AI bubble is a calculation by Silicon Valley venture-capital firm Sequoia that the industry spent $50 billion on chips from Nvidia to train AI in 2023, but brought in only $3 billion in revenue. That difference is alarming, but what really matters to the long-term health of the industry is how much it costs to run AIs. Numbers are almost impossible to come by, and estimates vary widely, but the bottom line is that for a popular service that relies on generative AI, the costs of running it far exceed the already eye-watering cost of training it. That’s because AI has to think anew every single time something is asked of it, and the resources that AI uses when it generates an answer are far larger than what it takes to, say, return a conventional search result. For an almost entirely ad-supported company like Google, which is now offering AI-generated summaries across billions of search results, analysts believe delivering AI answers on those searches will eat into the company’s margins.

In their most recent earnings reports, Google, Microsoft and others said their revenue from cloud services went up, which they attributed in part to those services powering other company’s AIs. But sustaining that revenue depends on other companies and startups getting enough value out of AI to justify continuing to fork over billions of dollars to train and run those systems. That brings us to the question of adoption.

Narrow use cases, slow adoption

A recent survey conducted by Microsoft and LinkedIn found that three in four white-collar workers now use AI at work. Another survey, from corporate expense-management and tracking company Ramp, shows about a third of companies pay for at least one AI tool, up from 21% a year ago. This suggests there is a massive gulf between the number of workers who are just playing with AI, and the subset who rely on it and pay for it. Microsoft’s AI Copilot, for example, costs $30 a month. OpenAI doesn’t disclose its annual revenue, but the Financial Times reported in December that it was at least $2 billion, and that the company thought it could double that amount by 2025. That is still a far cry from the revenue needed to justify OpenAI’s now nearly $90 billion valuation. The company’s recent demo of its voice-powered features led to a 22% one-day jump in mobile subscriptions, according to analytics firm Appfigures. This shows the company excels at generating interest and attention, but it’s unclear how many of those users will stick around. Evidence suggests AI isn’t nearly the productivity booster it has been touted as, says Peter Cappelli, a professor of management at the University of Pennsylvania’s Wharton School. While these systems can help some people do their jobs, they can’t actually replace them. This means they are unlikely to help companies save on payroll. He compares it to the way that self-driving trucks have been slow to arrive, in part because it turns out that driving a truck is just one part of a truck driver’s job. Add in the myriad challenges of using AI at work. For example, AIs still make up fake information, which means they require someone knowledgeable to use them. Also, getting the most out of open-ended chatbots isn’t intuitive, and workers will need significant training and time to adjust. Changing people’s mindsets and habits will be among the biggest barriers to swift adoption of AI. That is a remarkably consistent pattern across the rollout of all new technologies. None of this is to say that today’s AI won’t, in the long run, transform all sorts of jobs and industries. The problem is that the current level of investment—in startups and by big companies—seems to be predicated on the idea that AI is going to get so much better, so fast, and be adopted so quickly that its impact on our lives and the economy is hard to comprehend. Mounting evidence suggests that won’t be the case.

Conversation691 Comments By joining the conversation you are accepting our community rules and terms. To view rules, terms and FAQs, click here. Questions can be sent to moderator@wsj.com.

Conversations on news articles and news columns must be kept on the topic of the story. In opinion content, conversations can delve into other subjects. The conversation will close on this article four days after publication. What do you think?

Sort by

Newest

Kevin Walmsley 55 minutes ago The reason the Big Tech companies are tripping over themselves to find new billions to pour into AI is because of how it is going to replace the highest value labor. The WSJ just did a piece a few weeks ago about how even recent computer programmer grads are struggling to find work because AI can do most of the basic programming.

I remember when the naysayers gleefully wrote about the death of the internet when the tech bubble collapsed. Today’s iteration sees woke-inspired black Vikings or our constitutional convention attended by Indians (with dots, not feathers, no less) and say that AI is already done. Within 15 years, AI will replace most of these people working in these fields: pharmacy, riverboat pilots, mechanical engineering, radiology, geology, insurance billing, portfolio management and bond trading, materials sciences, aeronautics, pharmaceuticals, and dozens more. A robot that replaces a minimum-wage teen flipping burgers is worth $15 an hour. One that replaces a radiologist is worth $4,000 a day. And Google wants that $4000 a day, times half the developed world’s radiologists–plus everyone else in that list above.

Reply ·

Michael Goldman 2 hours ago This piece will not age well.

The writer is only talking about the most general purpose Chat AIs. Some narrow purpose AI implementations are extremely cost effective.

One therapist I know said it saves him and his company’s other therapists many hours a week in writing up the paperwork for their clients’ insurance cos..

Another at a giant insurance co. uses Ai to figure out what policies are most cost effective for their clients and the co. given the many legal issues.

A doctor I know uses a narrowly tailored AI to search the literature on specific medical problems he deals with. The list goes on and on.

AI has just begun. Add in humanoid robots to do the dangerous and monotonous tasks currently done by humans and it will be at least as transformative as the internet - probably more so.

Reply ·

Arne Christensen 3 hours ago Elon Musk is known to have a loose relationship with the facts. And, why should we expect Nvidia, for example, to downplay the abilities of its chips?

Reply ·

ALAN SNOWDEN 5 hours ago There is a very important distinction, which investors should never forget, but usually do get confused about. That is the difference between investing, and speculation (also known as gambling.) Investing involves buying a stake in a business with an established and time-tested business model, and a track record of positive cash flows. Speculation involves buying a stake in a business that MIGHT be the next big thing, and MIGHT become a successful business model, and MIGHT produce a flood of cash flow, but which very often turns out to be a cash eating, black hole. Those who can always remain clear minded about whether a business venture is investing or speculation will avoid many losses, and a great deal of remorse and embarrassment. Those who think they are “investing”, when they are actually speculating, will usually regret it later (and only occasionally hit the jackpot. ) (Edited)

Reply ·

James Bland 6 hours ago 12:59 , December 1999. Yawn this the best stock short of the century. As this article so proficiently pointed out , Bogus revenue AI apps are dropping from the sky. Most of the folks typing here (maybe) saw the tech bubble pop. But the promise of a “ terminator”! Or a colossus :the Forbin project. Not yet

Reply ·

1 ·

SIV T 6 hours ago I run business transformation practice for my company. What I see as key issue is not the technology - it’s the lack of clarity on the business problem or opportunity for which AI needs to be applied.

Reply ·

1 ·

Jake Barnes 7 hours ago I think everyone is missing the main point of the AI Revolution.

Sure, in the beginning, the benefits will go to the chip makers. But the real revolution is going to happen at the individual Business Level.

When Target, Macy’s, and Amazon merely embed AI into their inventory management, they will drastically reduce costs and increase productivity. And put any Procrastinating Competitors out of business.

Those insurance Companies, Banks, Auto Repair, Retail store inventory management, Etc. will be reaping the whirlwind, while competitors wither and die.

When “some” Insurance companies deploy AI, Drilling right down to who is driving, day, night, age, speed etc, 75% of Insurance rates will drop to the 85% that have no accidents. And put their Procrastinating Competitors out of business. It’s all the businesses that can figure out how to make the great AI leap in cost savings, massively increasing productivity, and give customers what they want when they want it —those are the ones you want to invest in today. And yes, eventually, today’s actual chip makers will likely be mere commodities.

Reply ·

3 ·

Jerry Harris 7 hours ago I think it’s safe to say, we have no idea what is to come with AI. Good chance no one has thought of, yet, what the technology can/will do. Just like the beginning of the internet.

Reply ·

3 ·

alan reyes 8 hours ago AI as a tech is a different question than to how AI will monetize as a set of corporate stocks. Unquestionably, AI can do things and interface in ways that open new industries. Money making and stock values don’t correlate now to actual industry function.

Reply ·

tim mccloskey 8 hours ago There’s no question about the amount of hype AI has generated and the massive investments required to just start to see benefits. But this Christopher Mims piece is nothing short of a bag job. Is there any acknowledgement of the far-reaching potential of both Generative and LLM AI? It will take gobs of money and lots of time and patience (with plenty of hiccups along the way) but the capacity for growth and limitless applications should at least be mentioned in a story under the rubric of WSJ Technology analysis, reviews, advice. How about a quote or two from developers, coders or just promoters? Naa. Let’s just bad-mouth it.

Reply ·

2 ·

Patrick Bonfils 9 hours ago If canards like Crypto, AMC and GameStop can be hyped to valuations in the billions, it should come as no surprise that AI is being overhyped.

But in this case, AI will at some point probably take over our world. When? Who knows. Can you invest to profit from that? Be careful, you just might be buying GameStop.

Reply ·

Patrick Bonfils 9 hours ago My hope is that the next big AI accomplishment is to replace the US political parties with something effective, efficient, and much less costly to run the country.

Reply ·

2 ·

Ian Kucera 10 hours ago “The problem is that the current level of investment—in startups and by big companies—seems to be predicated on the idea that AI is going to get so much better, so fast,”

This sounds suspiciously like the dot com bubble of the late 1990s. The internet and technology certainly revolutionized the world, it just happened about 10-15 years later.

Reply ·

Christopher Hoffman 11 hours ago The Russians build weapons. The Chinese build everything else. What do we build? Castles in the sky.

Reply ·

william Pottle 12 hours ago Fresh off his ‘what I got wrong about tech in the last 10 years’ the author has already written another banger that will be the centerpiece of his next lookback.

Reply ·

2 ·

David Lloyd 12 hours ago There has been an incredible amount of interesting comments so I won’t re-hash them.

As a person in the field I would say there are three things to keep in mind. First generative AI is new(er) unlike the uses of what we can call traditional or non-generative AI that have been commercialized actively for 25+ years, it takes time to weave it into the business fabric. Secondly, like all innovations, both chip /software based and not, they have had a measurable impact in changing industries (think PC, internet, and mobile for example) . Third, with generative AI specifically, the point is that AI won’t specifically eliminate your job, but the person who understands AI will. And that’s the norm for any change.

The largest issue with tech is that the hype rocket leaves for orbit way before it lifts off.

Reply ·

4 ·

Daniel Rice 9 hours ago I believe you make a controversial point that “AI won’t specifically eliminate your job, but the person who understands AI will”. Do you remember the early days of the PC when nobody could use them because they were based upon MS-DOS operating system and were so user unfriendly? In those days, it was only the rare person who understood that very complex and text based operating system that could use these computers. Only when Apple and Windows came out with very user friendly graphical based point and click systems did usability and sales dramatically increase.

Until AI is seamlessly integrated into its claimed applications in a similar way will it live up to the hype. But like MS-DOS back in those days, it has horrible usability problems and is very much prone to error – in this case caused by the software and not by the user. This suggests that those days are way off in the future and require different cognitive abilities than the very unintelligent parroting capabilities of today that often go awry just like parrots also do in mimicking humans.

Reply ·

2 ·

David Lloyd 3 hours ago Hi Daniel, I don’t think of it as controversial, and yes I was a software developer in those days. While usability is a very good point, I would ask whether you’ve used the newest OpenAI 4.0o model in web and specifically mobile. If you have then I would say the usability is rapidly improving. Concepts like prompt engineering to direct the way in which large language models respond is the “new” skill as was excel, access, ppt, and word in the 90’s and 2000s. Those that picked up these skills and knew how to apply them accelerated more quickly. Where I would make the argument in the shorter term is the accuracy of these models. Individuals, for example lawyers, using LLMs need to ensure they review the results and must have a level of expertise to ensure the resulting accuracy from the generative solutions. But the ability for an LLM to learn abut contract law and then direct an LLM with a skilled lawyer to draft portions of or entire contracts is becoming feasible.

The difference here is that these LLMs sitting behind the user interface don’t require any O/S skills. This democratizes the use of AI for those that look to adopt it as part of the job they do. I see this everyday. So my argument is that all change takes time, but as technology has progressed it has greatly compressed the time needed to go from innovation to practical application.

Reply ·

Lisa Miniter 11 hours ago SO true! I also agree with point on the slowness of changing habits… albeit with a far more mature U.S. adult tech audience market (e.g. think mobile took sometime…then faster uptake on using iPads, etc granted this was primarily a B2C effort) which extends to the over sensationalized media… both in the trades/consumer sites/brands..

Additionally, agree with industry’s point on what MS and Google/Amazon are aiming to be a new market of ongoing revenues…via a wide swath of co’s/startjups to continue to use their respective AI / cloud computing offerings.. (e.g. w/ the biggest tech players on the planet…that in itself is a core point that I am not aligned with until more of this plays out.

Reminds of 3D content back in 2010 …the massive media with mobile, TV, movies - was way ahead of reality and didn’t’ work due to lack of content (aside from entertainment/sports) and the tech infrastructure needed to efficiently scale at lower costs to make $$ from ongoing audience revenues. (Edited)

Reply ·

1 ·

David Lloyd 11 hours ago Completely agree, with one exception that new venture backed companies will flame out in masses, it’s the nature of how industry transformation works. It’s always interesting how 80% of the largest companies from 25 years ago aren’t anymore. Some re-invent, but creative destruction has a way of catching everyone over time.

Reply ·

Lisa Miniter 10 hours ago YES….like with GE from 25 years ago…and GE’s “creative destruction:” when they lost power starting in ‘09… with its massive divestures of disparate “creative” entities…remember NBCU and their commercial RE division, ..recall its now run as 3 co’s I think under the umbrella term of a “GE Company”…but could be wrong…your point is well taken. (Edited)

Reply ·

1 ·

B Walker 12 hours ago This all makes a good deal of sense, and accords with what I have been experiencing. Also, it makes students think that all that matters is the answer – not the critical thinking process. So they cheat, and as a prof I now have to make them hand-write all their papers in class, with their electronic devices at the front of the room. (Edited)

Reply ·

2 ·

petey b 12 hours ago Generative and LLM AI is for sure being hyped. I am not sure I believe that the situation is as bad as claimed. The main value for LLMs and applications of generative AI appears to be in the ability to understand a question, and form a coherent response. That has value, but is not revolutionary. It’s an improvement on previous interfaces for search. It’s revolutionary only in the sense of how much better it is in acting like a human than prior art. But it’s only evolutionary in terms of the actual economic value.

Reply ·

Lisa Miniter 11 hours ago Agreed, though it seems with your point of AI’s “ability” to understand a question….

That’s where the perceived “value” of this task/search is pre-formatted orengineered with feedback (at scale) that will make “critical” thinking less appealing to the Gen Z sector in particular…

Obviously not all kids…but am certain the overwhelming majority will be “trained” and influenced via (daily repetitive usage of say MS Co-Pilot) to take short cuts from what has been stated across credible acadamia voices for years… .the lack of critical thinking due to massive reach of these search conveniences

The undercurrent of this is happening now via searching cover letters, resumes, proposals and likely tweaking accordingly…

In the aggregate of this AI usage, seems to negate the “value” of critical thinking for this very reason…my opinion however its being played out now with Co-Pilot and others.. (Edited)

Reply ·

Jerry Mclaughlin 12 hours ago After a deep dive into developing a research app to do one simple thing use ChatGPI to reverse itself create questions and then deliver 4 wrong answers and one right answer without duplicating itself. It is almost impossible. Somewhat like mining for gold, a huge amount empty pans. In simple coding tasks it will, between iterations, forget even when told what worked, repeat the same error, and then always add more complexity even when told to simplify. The danger with AI is that people will treat it like a calculator and assume it is always right. For AI to expand the chips will need to solve the heat problem (not just figure out new ways to suck the heat out) so that its growth is not tied to the production of electricity; and secondly offer or incorporate 2nd level responses that anticipate the next problem. They are completion engines for one off questions but not for solutions.

Reply ·

Thomas Worosz Jr. 12 hours ago Worked in IT for 40 years coding, designing/installing large systems (data bases, applications), and finally selling them. As I retired data bases were commoditized into business applications and data mining into analytical tools. When one googles AI, descriptions lean towards applications that are self learning, neural networks that mimic human thought, unsupervised learning, autonomous operation - basically give AI the problem and it will create the solution/answer/system. Like the evolution of languages (COBOL, FORTRAN, PL1, C+), data base systems(DB2, Informix, ORACLE, SYBASE) or system design tools (PERT, CAD/CAM, GUI, INTERNET)…AI needs to evolve to be proved useful in terms of functionality/value. Utilize AI to develop fundamental applications in accounting, government, medical professions - mundane tasks that require pouring/analyzing/recording data/results. Look at our bloated government bureaucracy - IRS analyzing tax returns - a plethora of data; medical profession where doctors/nurses spend their valuable time being data entry clerks. AI, like other evolving technologies, needs to take baby steps before it can go out into the world to prove its worth. An old timer’s view…have a great day.

Reply ·

2 ·

Jerry Harris 7 hours ago Agreed. Smart, driven people will find good use of this technology, in time. Still in the 2nd inning. Long way to go.

Reply ·

Lisa Miniter 11 hours ago Your definitely not an old timer…

Reply ·

Alan Moffatt 12 hours ago Wrong. AI is destroying the ad agency business with more clients taking more advertising in-house to use Al for strategy, media and creative development. Inflation has exacerbated this because COGs have risen so companies need to cut expenses (i.e. agencies).

Reply ·

Edward Dunn 12 hours ago Gen AI is also destroying stock photo and stock media firms that drives these same ad agencies - the client can use DALL-E to create the images for their marketing

Reply ·

David Ragland 13 hours ago Complete balderdash! I’ve had a decades long career in technology, and currently own a small AI consulting business. To date, we’ve written more than 100 empirical white papers on GenAI, and have built a functioning fine-tuned model, etc. etc.. With all due respect to Mr. Mims, had he done the same, he would not have penned the above article. I say this NOT as someone who is a big fan of GenAI –on the contrary, If I could get in a time machine I’d go back to the 19th century (and would only take penicillin). I dove headlong into GenAI more out of concern for humanity. We are currently living in-between ANI and AGI., and much much more is coming (soon). Deep-neural networks, ReLu’s, sigmoid functions, and the “black-box” interpretability mysteries they help propel should be enough (alone) to dispel the position of the above article–but there’s SO much more happening–and it’s happening fast! (Edited)

Reply ·

1 ·

Jerry Mclaughlin 12 hours ago Really impressive technical jargon but what real life value based solutions do your white papers demonstrate AI producing? I would love to read your white papers but I understand the limitation of these comments in not allowing links . Not getting personal but your profile on your WSJ image needs to use the first AI we had and realize that retired is not the same as retried. (Edited)

Reply ·

1 ·

David Ragland 9 hours ago Once again, the media has led society astray (they’re capitalizing on helping us deal with cognitive dissonance–which sells, but isn’t accurate). Here are 7 practical uses just off the top (for good or bad). 1. Content Creation (this one’s obvious), 2. Healthcare - Drug Discovery: AI models are used to generate potential drug compounds. Medical Imaging: Generative models enhance medical imaging techniques, helping in the creation of high-quality images from lower-resolution scans. 3. Entertainment- AI is increasingly used by artists to generate new compositions and assist in the creative process, Film and Animation (not that I like this and many others); 4. Design and Fashion generating new clothing designs and predict fashion trends (not to mention Product Design in engineering and manufacturing .5. Finance- Algorithmic Trading: AI models analyze vast amounts of market data to predict stock movements and execute trades, as well as Fraud Detection: 6. Customer Service— virtual assistants like Apple’s Siri, Google Assistant, and Amazon’s Alexa. 7. Education - Personalized Learning: AI systems generate personalized learning plans and educational content based on individual student needs and progress. (Edited)

Reply ·

2 ·

Lisa Miniter 11 hours ago SO true! Meanwhile the largest tech giants on the planet .e.g Oracle, MS, Google, Amazon…are the one’s in the driver’s seat and CAN outspend anyone in the long game…the “question” is how recurring will the revenues be….likely starting by 2027 and onward…from the co’s/startup clients who would use their “valued” Gen AI/cloud computing to prove ROI …branding like TUMI luggage etc…. will at that point continue to be far more diluted…in its persona…we’d be far too tech commoditized across lots of goods/services at that point …granted above are leading brands…but the biggest brand in AI…NVIDIA (sp?) is bleeding cash…and overvalued. (Edited)

Reply ·

C HULSE 13 hours ago Check out the May 13 WSJ article - “Suddenly There Aren’t Enough Babies. The Whole World Is Alarmed. “Birthrates are falling fast across countries, with economic, social and geopolitical consequences”.

It’s talks about the looming The Demographic Winter. My conclusion is that the workforce is going to continue to shrink while societal needs will remain and/or perhaps even increase (more and more retirees, people living longer, etc.). That’s why I’m bullish on, and choosing to embrace AI as it will be necessary to perform many of the existing jobs in the near future.

The article seems to focus on having consumed what’s available on the internet and fails to mention the vast amounts of business and customer data, robotics, programming, etc that AI will also replace.

Reply ·

1 ·

Reid Cuming 14 hours ago Anything bad for AI is good for humanity.

Reply ·

Peter Deserto 14 hours ago The army with the best technology always wins the war. Companies that can use AI effectively in operations will have a significant advantage over their competition. This technology is in its infancy. These are exciting times.

Reply ·

Edward Dunn 12 hours ago Somalia, Korea, Vietnam, Iraq, and Afghanistan are clear examples of debunking the “best technologies win wars” line of thinking.

Reply ·

Mark DeBruler 14 hours ago I’m not sure AI is losing steam so much as people are trying to figure out how this new tool can best be used. The author seems to forget (or wasn’t around for) the Dot-Com bubble, when companies were trying to figure out how the run a profitable internet company. Many tried and failed, but look at today!

Reply ·

3 ·

Jerry Harris 7 hours ago And a lot of those were startups. Today, with AI, it’s all the big dogs with a bunch of money sitting on their books. They can wait out the technology advancements coming.

Reply ·

Mark DeBruler 5 hours ago Good point!

Reply ·

Arthur Fisher 15 hours ago As with almost anything else, the question is what will it do at a cost that people can and will be willing to pay, and how many such folks are there? A second question, implicit, is this: Is this the short-sale of a lifetime? It will be interesting to see what Apple, always very sensitive to issues of ease of use, but also very susceptible to the techie temptation to over-complicate, brings to this circus. Plainly not all of the related costs (training time, electricity costs, copper, etc.) are yet figuring much into the calculations.

Reply ·

Thomas Bradford 16 hours ago The reality is that AI has been overhyped by investors who are always looking for the next Iphone, and by a MSM that will write story after story to back it up. because their business is hype. But, as always, such innovations are far more slowly integrated into a world of commerce than advertised. The people who actually benefit are those who begin the new start-up, pay themselves enormous sums or pocket scores of millions from premature IPOs, and then watch as it fails and all those who invested lose their shirts. The same is happening with far too early AI based startups that make promises they can’t possibly keep. And the same suckers…. thousands of small individual investors with no access to real knowledge about their investment….. keep buying into it.

The smart money comes from a small inner circle of investors who quickly begin to exit the investment as the money pours in to recoup their original investment, and then watch carefully to see if they should let their earnings ride or pull out before everyone else understands what is happening.

Reply ·

Alan Johnson 16 hours ago AI, climate change, COVID, tulip bulbs ….. all overhyped. People get excited.

Reply ·

George Hresko 21 hours ago In just a few words how does AI handle the johari window?

Reply ·

D L 15 hours ago From ChatGPT: AI can help facilitate the Johari Window process by analyzing input from self-assessments and feedback from others to categorize traits into the four quadrants: open, hidden, blind, and unknown. This can aid in personal development and improving self-awareness.

Reply ·

1 ·

Mike Wracher 21 hours ago It’s amazing to me that a human brain running on a ham sandwich and some potato chips perform so impressively, and it only took 300million years of mayhem to create.

Reply ·

8 ·

Jerry Mclaughlin 12 hours ago Yes I use the oreo cookie fuel. We obviously over that time period solved the heat problem (I consider quad espressos afterburner)

Reply ·

Hans K. 22 hours ago “The pace of innovation in AI is slowing, its usefulness is limited, and the cost of running it remains exorbitant”.

Good, now let’s return to real life!

Reply ·

2 ·

Dan White 22 hours ago For 30 years I respected WSJ as gospel. Poof. Wrong Wrong Wrong.

Reply ·

3 ·

Neven Karlovac 23 hours ago In my experience, Perplexity from Anthropic doesn’t hallucinate and always provides sources… two essential features for a trustworthy service and serious use.

Reply ·

2 ·

Georgi Stoyanov 21 hours ago Agree about Perplexity, it’s great for research as it gives you sources you can verify yourself. So even if it hallucinates it’s much easier to catch. I think Perplexity is an independent (private) company though, not affiliated to Anthropic, no? The free versions uses OpenAI’s GPT-3.5 as a model while the paid one gives you a choice between multiple models, including GPT-4 and Anthropic’s Claude.

Reply ·

2 ·

Neven Karlovac 12 hours ago Thank you for the correction. I asked Perplexity itself who owns it and they answered the question. The company is owned by three co-founders, all techies with AI background and and a number of VCs and individual investors. As indpendent as it gets.

Reply ·

Benjamin Fairless 23 hours ago Payroll costs could be greatly reduced by ai assisting in CEO functions

Reply ·

2 ·

ERIC M ELLIS 17 hours ago That’s really funny. Not sarcasm. I mean it wholeheartedly. I can’t believe how many CEOs risked burning off half their customers to go “woke.” It might be the biggest tell in modern history.

Reply ·

3 ·

Matthew Umbel 23 hours ago “AIs still make up fake information” - hahahahaha, just like Elon does.

Reply ·

Joan Newcomb 16 hours ago And Lina Kahn.

Reply ·

Ackley Wang 1 day ago 10 years ago, self-driving cars were super impressive and looked to be the way of the future. I told my kids they probably wouldn’t need to bother with a driver’s license.

Hahaha, How wrong I was. My kids need those licenses. And the self driving capabilities have barely improved at all.

I do believe we will get there but it’s clearly going to take a LOT longer than the fanboys claimed.

Looks like AI will follow a similar trajectory as Nvidia rakes in a huge boatload of cash from AI investors.

Reply ·

3 ·

Frederick Scholl 1 day ago I think AI will be imbedded in everything and also lead to break throughs. Too early to write it off.

Reply ·

1 ·

Charles Stout 16 hours ago Break throughs probably true, BUT on a much slower timescale than currently hyped. (Edited)

Reply ·

Juan Roman 1 day ago So far the only thing AI has done for me is get annoyingly in the way when I try to search on the internet or call customer service to resolve an issue. Nothing like spending the first 10 minutes navigating AI on the phone to be then transferred to an overseas answering service for an hour before finally reaching a person in the US to reschedule my cancelled flight.

Reply ·

3 ·

Anne Bennett 21 hours ago Calling big companies has been nightmarish for a long time. Has it deteriorated even more?

Reply ·

Gary Hilton 1 day ago Did Kathie Woods write the article?

Reply ·

1 ·

Rahul H 1 day ago Odd Comment. She’s on the opposite side of the case (Cathy).

Reply ·

1 ·

Mark Stamp 1 day ago But significant disappointment may be on the horizon, both in terms of what AI can do, and the returns it will generate for investors.

Bingo!

Reply ·

3 ·

john boyer 1 day ago Artificial Intelligence is based on flawed inaccurate information fed to it. We have enough problems now without relying on AI. The genius of humans is we make mistakes until we get it right enough to be useful.

Reply ·

1 ·

Paul Leavitt 1 day ago First there was the “metaverse” and now there’s AI. I wonder what the third thing is that is soon arriving and is about to change our lives forever and require billions of dollars of new investment posthaste.

Reply ·

5 ·

Harrison Siegel 1 day ago This article fails to explore how generative AI is adding tremendous value in the fields of medicine or healthcare (specifically cancer related research, pattern identification and machine learning from health care records), or integrations with communication technology such as Teams, which now can transcribe meetings, summarize the main highlights and takeaways, and help the participants and others absorb the content later. This technology is just starting to be utilized for a variety of productivity related tasks and is going to continue to accelerate as humans think of other ways to apply it in novel ways.

Reply ·

8 ·

Ackley Wang 1 day ago Transcribing and summarizing meetings? The always leads to major scientific breakthroughs, right?

Reply ·

1 ·

Jeffrey E Rothman 1 day ago This layman sees it as google on steroids. It can go through more data and perform more mundane tasks more quickly than we were able to do pre AI. This may pour over more data more quickly at a medical office. But ultimately, people have to use their analytical skills to benefit from its results

When there are specific applications of it my field, I’ll pay more attention to it.

Reply ·

1 ·

Joe Doe 1 day ago We need more articles detailing use cases for AI. Generating pictures and text is just one small aspect of the technology’s potential. I want an AI radiologist ASAP. I don’t want AI search results as they are just search results packaged into Progressive sensibilities. AI will do great things for oncologists as they are never completely up to speed on the latest research. When AI can read all of the relevant research and make treatment recommendations, along with providing the prognoses for different therapies, people will benefit. The same goes for diagnostic work. Doctors get used to ordering certain tests because they understand them. AI should be able to optimize decision-making so diagnostic testing is done in the most cost-effective way possible. Lots of things for journalists to cover.

Reply ·

6 ·

Alan Nathan 1 day ago I use AI extensively to run complex monte-carlo simulations , while simultaneously generating the relevant python code to share with others. This saves me tons of time and I’ve checked the math on the simulations vs the old way. I also use it to get back of the envelope calculations for bond YTM & YTC. It does have a bit of an error band around it as has probably been noted elsewhere, but it is trainable and it does error correct when errors are pointed out. So I’d say it speeds up process for those with subject matter expertise, and it absolutely improves my abilities to form useful queries.

Reply ·

5 ·

ARTHUR GANDOLFI 1 day ago It’s still just a bunch of algorithms. All it can ever do is copy existing sources and reassemble it.

We just tried to use AI to categorize some free form text at work. It missed THE single most obvious trend. I would expect better from a dullard intern. (Edited)

Reply ·

1 ·

Miles kirkhuff 1 day ago I’ve been seeing a lot of AI Biden political ads on YouTube. He looks so energetic and annunciates his words so you can definitely tell its not real.

AI can do some cool things but its still got some serious work to do.

Reply ·

4 ·

Larry Pickett 1 day ago You can’t judge the outcome of a baseball game in the first inning. This is the first inning of Generative AI. It will get progressively better and every white collar worker will use it like they use the phone and internet today.

Reply ·

1 ·

Howard H 1 day ago Mims will regret writing this article. And not in the far future.

Most of the white-collar employees in this country are employed in jobs where they serve as custodians, executors, or advisers on business processes. Think attorneys, procurement, managers, supervisors, CSR’s, order processing and fulfillment, design, programming, engineering, research, QA/QC, logistics planning, accounting, billing, etc. EVERY SINGLE ONE OF THOSE PEOPLE ARE AT RISK OF BEING REPLACED IN TEN YEARS. That’s the corporate holy grail of AI. 70%+ reductions in employees to provide the same service (or perhaps better, looking at you Xfinity).

Is it expensive? Of course, especially early. But an AI system doesn’t take sick leave, need a pension, or file discrimination lawsuits because its server center wants a purple top instead of a blond one. It also (probably) doesn’t steal IP and go work for China. Hopefully.

Where automation and robotics transformed production over the last 30 years, AI will transform “knowledge workers” and not to their benefit. It will be faster and captive — at least now — to electrical generation capacity, but it will be a change comparable to the Industrial Revolution in a fraction of the time.

Blue-haired baristas will be out of luck.

Reply ·

8 ·

Scott Johns 1 day ago If all those professionals are going to be replaced by AI then how will those professionals make income to support themselves and their families.

Is it ethical to continue developing AI if it will result in a 25% unemployment rate?

Reply ·

2 ·

Al Min 1 day ago If those are professionals they will use new tech as a better shovel to do their job. One-trick devops ponies are not professionals. They will be freed from mundane tasks to apply their creativity elsewhere.

Reply ·

Howard H 23 hours ago I’m looking forward to the day when AI’s are generating meaningless emails and exchanging them with other AI’s on issues of no import to accomplish little if anything, except at the TB/sec rate.

I call it Dilbert’s revenge

Reply ·

2 ·

Morgan Masner 1 day ago Just wait, what you see now is nothing and AI doesn’t follow Moore’s Law, it’ll come like a tidal wave.

Reply ·

1 ·

Michael Flanigan 1 day ago Just like self-driving cars did. /sarcasm

Reply ·

2 ·

Kishore Dandu 17 hours ago funny!! as an enterprise data engineer, we are hardly keeping up with clean and reasonable data availability .. this article did not cover that aspect of bunch of data issues in big enterprises that won’t go away magically, just because some one perceives ultra futuristic use cases!!

Reply ·

Cory Pratto 1 day ago “Trust me, bro.” - Some Guy on the Internet

Reply ·

5 ·

Margret Martin 1 day ago There is a lot of hype, but there is an amazing amount of innovation that has been coming out over the last year. . If it wasn’t so powerful the government would not be restricting access to it

Reply ·

Daniel OBrien 1 day ago I am surprised by the comments. Normally a more educated crew. I guess all the doubters feel safer with Their money under the mattress.

Reply ·

2 ·

Scott L 1 day ago All my money is in pets . com and Webvan

Reply ·

JR Katz 1 day ago This column is comical. The personal computer were also exorbitant when they first came out. As technology progresses the costs always go down and speed and efficiency rises. This is always the case. It’s ALWAYS been the case as well with AI since it started.

Reply ·

3 ·

Deepinder S 1 day ago US fighter pilots that are losing to AI flown jets might disagree with this assessment. It is only a matter of time when businesses become as proficient at applying AI as the military is. Sure AI makes mistakes, but what has been accomplished is already giving hiring managers pause across all industries. There might be an AI blip or a correction, but one cannot be glib about AI’s potential outcomes. (Edited)

Reply ·

3 ·

Matthew Minor 1 day ago But, but, Elon said AI is going to take over the world?

Reply ·

James Crawson 1 day ago Several big misconceptions in this article. And the biggest wrong assumption of them all is that what is coming down the pipeline is going to use the same architecture of today. It continues to surprise me that 99% of even educated persons are unaware of the breakthroughs in new AI architectures yielding shocking results. But hey, ignore AI if you like, more opportunities for me.

Reply ·

3 ·

Al Min 1 day ago Besides LLM and image generation, what do you classify as AI which does not fit into old school ML?

Reply ·

Karl Wilcox 1 day ago We live in a fantasy driven society where tech has achieved the luster of godlike powers. Perhaps investors should consider Icarus.

Reply ·

3 ·

Gerald Edwards 1 day ago Agreed. Technology is a good slave but a bad master.

Reply ·

3 ·

J Gault 1 day ago AI is just another fashionable marketing buzzword, like sustainable, green, etc.

Reply ·

4 ·

Sami Harawi 1 day ago The article is very worrisome about investing in AI. It comes back-to-back to that James Mackintosh wrote for the WSJ on May 30. Both are fascinating eye openers for AI investors. Is their pessimism about AI going to translate in a routing of the Markets on Monday? One thing that neither articles address is INFERENCE, meaning how the trained hardware will create a new intelligence extracted from the data fed. I do not know myself enough about that to give a definite opinion, but it seems to me the place innovation will flourish there. Also, neither article mentions some established uses of AI. It seems useful in various aspects of health care and possibly in Quantum computing. Thank you both for cautioning me about AI investment

Reply ·

1 ·

Dennis V 1 day ago Tbh we shouldn’t call all of these new apps “AI”. The term is being misused. For example, rendering an image and up-scaling, is now called AI. Nonsense.

The term has lost its meaning. And most “AIs” out there are more advanced algorithms, nothing more.

This is all about raising funds

Reply ·

3 ·

John Newman 1 day ago One of the first things I did when I set up a free account with Chat GPT was ask it where the best place to go in Tampa to find a streetwalker was, and where I was most likely find someone selling crack cocaine. It wasn’t because I wanted either a hooker or drugs, but I wanted to see how filtered the information the tool provides was. A lot, it turns out. Both times I got moralistic lectures about looking for things that are illegal. So right off the bat, that makes today’s version of “AI” seriously flawed and next to useless.

I do use it occasionally at work for queries like, “give me an AWS cli command that, given a private ip address, will find the instance associated with and tell me the volume id’s of the EBS volumes attached to it and whether the DeleteOnTermination tag for each is true or false.” And it will generally spit out something that doesn’t quite work, but is at least close enough to working that a few minor tweaks gets it working. And this does save me some time over formulating it on my own.

I also usually tell it what fixed it, and although it thanks me, I don’t know how much impact that has on its “learning”. Probably some, because it has slowly been getting a little better over the past 18 months. I also notice that Chat GPT is far superior to Microsoft CoPilot on stuff like this.

Reply ·

Daniel OBrien 1 day ago I’ve used Gemini to compare ETF’s for example and the results far exceeded my expectations. Advanced is pretty remarkable. Garbage in gets garbage out.

Reply ·

1 reply

Matthew Zobian 1 day ago It doesn’t seem like AI has the economic advantage of “network effects”, unlike, say, operating systems, search, and CPUs and GPUs. This is another reason why it will likely commoditize to at least some extent. Nobody will dominate AI like Google dominates search or Nvidia dominates GPUs.

Reply ·

1 ·

P Z 17 hours ago It will be a market share battle: the winner will sign up the most paying customers to provide recurring revenue stream to continue investing in their AI products.

Customers will decide which AI product will be the most objective, the most up to date, the most accurate, the fastest, the most convenient, the most integrated with their own personal data to help them with their tasks, etc.

Whether or not the general public will want to pay for AI services is yet to be determined, because historically these kinds of services have been “free” and paid for through advertising revenue. As AI improves and becomes more indispensable for daily life, users will likely see the advantage of paying for the most features (like upgrading seating/perks on flights).

Reply ·

Eric Porter 1 day ago Like the dot com bubble, there will be a bust where many “AI” companies go out of business. Many of them aren’t really doing anything new but calling it AI to attract investor attention.

Amazon stock fell over 90% but came back stronger. Maybe AI can power a new company to a trillion-dollar valuation, but so far the only real winner is NVidia, the company selling picks and shovels to the gold miners.

Reply ·

7 ·

Gerald Edwards 1 day ago Truer words have rarely been spoken on this topic.

Reply ·

Steve Anderson 1 day ago The hype on any tech rarely matches the reality…and if it does it’s only after a longer period of time

Reply ·

Robert Rainish 1 day ago Already articles about the demise or will not be very good regarding AI. The infrastructure is only now being built. The application software especially for the enterprise is now being written. The databases to use be are still to be created. The only software we see being used is consumer programs that are in testing stage. The only problem is people expect a finished product today rather than understanding it takes time. Anybody remember Netscape, one of the first internet browsers. AI is transformative. Will take time and will be as wealth creating as the iPhone or Internet or Microsoft Windows.

Reply ·

2 ·

Albert Zhou 1 day ago No technologies ever changed the society as fast as these promoters claim. The PC started in early 1980s and most people do not use them often until at least 2000, so expect the take-up rate for AI to be very gradual and limited in a few domain where labor cost is exceptionally high, hopefully in healthcare industry.

Reply ·

1 ·

Dan M 1 day ago AI is really good at generating computer code and songs. This was true with GPT 3.5 and hasn’t gotten better since then.

Reply ·

Ray Loehr 1 day ago Ai is also being misused. Recently a few kids put pictures of real people on fake naked bodies. The affected teenagers were deeply hurt and faced frequent ridicule. I am dubious of nearly every photo I see.

Reply ·

Cory Pratto 1 day ago You could do that with photoshop too. AI just made it easier for anyone without photoshop skills to do it. You could even make fake videos before AI, not just pictures.

Reply ·

1 ·

Albert Zhou 1 day ago Yes, you should be very suspicious of any pictures and videos on the Internet.

Reply ·

Show 1 more reply

J steffy 1 day ago The AI frenzy will run its course ; similar to the Dot.com bust of 2000. There will be a massive sell off of NASDAQ related AI stocks by this August. I predict a minimum loss of value in these equities of 25%

Eventually the strong AI companies will grow in value again but the near term downturn will take the entire S& P down as well. There will be some very good value shopping at that point, but it will take AI related stocks a couple years to come back to their present high valuations.

It will take another couple of years before companies and the public see any true benefits to AI innovation. It’s 80 % hype and 20% reality.

It’s going to be very stormy market soon. so it’s best to buckle up. (Edited)

Reply ·

ian board 1 day ago One of the most interesting curves, that you see examples of everywhere, is the logistic function. At the beginning, it looks exponential, in the middle, roughly linear and far enough out, constant. It describes things whose growth is limited by some key resource - why the grass on your lawn will never get 10 ft tall. Given the electricity demands of AI and the specialized chip making (really only one fab - TSMC) I would expect the AI boom to level off fairly quickly. We’re probably already somewhere in the mid-region of the curve for AI.

Reply ·

Rita McGrath 1 day ago Whether disruption is a thing or not depends very much on how long you give it. While in the near-term, established companies do a great job of defending themselves, research reliably shows that even mighty companies like the Fortune 500 top out at about 40-50 years on average before they are split up, acquired, go out of business or disappear entirely. I think Christopher Mims’ observation about moving faster is spot on, though. What’s the answer? Watch for my next book!

Reply ·

1 ·

John Mellen 1 day ago Rita, your book “Seeing Around Corners” (available on Amazon) looks interesting. 📚

Reply ·

Matt Shnark 1 day ago The easy targets for AI are things like education, which are unsurprisingly protected by many layers of political and bureaucratic shielding.

Ironically, entry of AI at scale into education might provide the kind of cognitive gains at scale which can lead to more uses for AI.

But when natural intelligence is capped by a dumb system, artificial intelligence suffers the same fate.

Reply ·

1 ·

P Z 17 hours ago AI education is like paying almost nothing for a full-time university-educated tutor for your children available 24/7. It is perfectly patient and never gets tired.

People aren’t talking about AI for education very much, and they’re definitely not talking about AI to replace federal bureaucrats and optimize code of federal law for consistency, accuracy, and brevity.

Reply ·

Albert Zhou 1 day ago Non-profit driven organizations have very little incentives to change anything, so I would have been very cautious about pursuing these vertical domains if I were working for a start-up company. Better focus on industries with deep-pockets.

Reply ·

1 reply

PHILIP J HELLERMAN 1 day ago Technology in search of an application.

Reply ·

1 ·

DAVID WINTERS 1 day ago It’s machine learning not “AI”

Reply ·

Cory Pratto 1 day ago Machine Learning is a subset of AI.

Reply ·

3 ·

Steven Eliscu 1 day ago This article provides a good dose of reality, especially for NVDA investors, who should be careful over the next year. A killer app for AI tools already is software development, the advances of which will be largely invisible to most. Companies who adhere to AI as a key tool to get the most out of their developers should get a huge boost. Software offshoring will get hurt, as AI should reduce that need, so it should be good for keeping developer jobs onshore. That in turn should help drive productivity in a way that will take years to ripple through the greater economy but should be enduring. AI is a big deal, but probably not in the way people think.

Reply ·

1 ·

Cory Pratto 1 day ago What companies are you talking about exactly as being AI killers?

Reply ·

1 ·

peter a cole 1 day ago AI will have the impact somewhere between the smart phone/the internet and the steam engine. We are way, way at the beginning of figuring out how to use AI and how it can be effectively put to use.

Reply ·

peter a cole 1 day ago The article indicates perhaps APPL is taking the best approach. That approach is “rent AI” to be used on Apple devices and provide the best way to access the best AI on Apple devices.

Google to maintain close to it’s dominate position in search has no choice but to invest heavily in AI.

Reply ·

Stephen Keith 1 day ago It seems to me AI is more incremental than revolutionary.

A sort of advanced means of making up phony content for the Internet.

Reply ·

ROBERT WEISBERG 1 day ago Thats what some will say without really diving in. A friend of mine developed an application which is as good if not better thna comemrcial products. Chat GPT wrote the python code. While he tells me he knows some phython, he says there is no way he could have completed it himself. AI will get better. It has made tremendous progress already. I have used it to help with excel and take my native files and apply complex code that if I could do it would take me many hours of frustration. That has not existed before.

Reply ·

1 ·

steve coon 1 day ago AI is just the latest “next shiny thing” pitched by snake oil salespeople making promises that too many gullible people fall for. There’s a sucker born every minute. This said, of course, there is a role for AI in today’s business, but it is not well-defined and therefore we see the problems associated with it. Gary Marcus quoted in this article is a wise entrepreneur who saw the truth and sold his AI start up. The story doesn’t detail how exactly the 75 percent of white-collar workers are allegedly using it. In my field of journalism, AI can be a useful tool for curating potential news sources and writing drafts of stories. But any legitimate journalist and editor would always confirm what AP provides—the technology hallucinates as has been pointed out in numerous reports. So, AI has a role but what is unclear, it is expensive, still required human oversight and will mean big money for only a few investors.

Reply ·

1 ·

Homer Luther 1 day ago Check the tech capex expenditures, check the record start up valuations, ask Mayo Clinic, check with tv and movie makers, check with people in our defense department and those banning or restricting NVDA chips to china and Middle East The article in the WSJ was click bait, its “evidence “ was unidentified research, a Wharton professor and the author’s suppositions laced with “if” “might” “could” etc — talk to people actually using it daily

Reply ·

3 ·

WILLIAM ROYCE 1 day ago I use Generative AI in a limited way to edit photos. It works great (most of the time) by reducing the time to do some tasks like colorize old images, or restore them. It also works well to eliminate objects and/or subjects and replace them with a matching background. I cannot see this technology eliminating photo editing jobs, however.

Reply ·

2 ·

ROBERT WEISBERG 1 day ago Give it time and many of the time wasting basic photo editing work will be automated fully.

Reply ·

1 ·

Andrew Arbenz 1 day ago It’s useful to see the bear case against generative AI, since there are so many presenting bull cases.

Reply ·

6 ·

Cory Pratto 1 day ago Watch out, you’ll get some schlub here to say you’re just sour grapes over missing the AI investing boat.

Reply ·

2 ·

2 replies

DAN WOJCIK 1 day ago Maybe we can get AI to replace Joe.

Reply ·

4 ·

peter a cole 1 day ago How about have AI fact check Trump every time Trump open’s his mouth or posts something on Truth Social or elsewhere? How about the same for when ever any of the MAGA operatives open their mouths or post on social media?

Reply ·

2 ·

2 replies

WILLIAM ROYCE 1 day ago You mean replace “NO I” with AI?

Reply ·

Alexandre Dniestrowski 1 day ago If you consider how long the internet (1995) or the smartphone (2010?) technologies progress unfolded, the bombastic assertions of the author are precipitated: let’s discuss these matters at the end of this decade and then “qui vivra verra!”.

Reply ·

1 ·

steve coon 1 day ago Alexandre, you are correct. Promotors of new technologies, or innovations of existing ones, always try to lure investors to their effort. It’s part of the capitalist practice and sometimes it works. But many investors have so much money that they can take the risk of raking in even billions more. When it isn’t profitable, they just move on to the next hyped offering. Thanks for your post.

Reply ·

DB Levene 1 day ago I think that training AI, which already suffers from hallucinations, on synthetic data will only increase the problems with AI’s unreliability. AI may have niche uses in areas where accuracy is unimportant.

Reply ·

5 ·

Anthony Brown 1 day ago The AlphaFold AI program predicted the 3-D structure of every known protein.

Reply ·

Cory Pratto 1 day ago Not a coder, but I’ve heard for coding it is very helpful. I’ve used it plenty for scholarly research type stuff. It is useful, but at best it is an advanced search query. Gets things wrong a lot even on very basic things. Not as revolutionary as people think it is.

Reply ·

3 ·

1 reply

Joseph Areeda 1 day ago I believe it was Donald Knuth in the 1950’s who said:

“The most amazing thing about modern computers is the uncertain grounds upon which we attach any validity to their output”

Still true.

I’m having trouble confirming source of the quote.

Links to this note

Artificial Intelligence