Month: October 2024

Categories
AI technology

Sam Altman (OpenAI) cannot stand in the shadow of Elon Musk

Just because the last few weeks were so often about Sam Altman of OpenAI and Elon Musk of Donald Trump, I was working on a newsletter with totally different topics. But then so much extraordinary happened in a few days that I still have to conclude that Elon Musk is underrated and Sam Altman is overrated.

Amid all the media frenzy about the rascals of the tech world, we would almost forget that two people also regularly featured in this newsletter, George Hinton and Demis Hassabis, have both won Nobel Prizes. Still, again, now focus first on Musk and Altman.

Musk and Altman opposites

The strange thing is that every mention of Musk evokes sharp reactions, with half disagreeing with me on principle because Musk is said to be a reactionary racist, while the other half find the achievements of his companies undervalued: either you hate the person, or you love his achievements. Whatever you think of the person; last week it became clear once again that Elon Musk ranks as a lone innovator. 

OpenAI is seen by most people, whom I conveniently describe as laymen, as a uniquely innovative company led by a mysterious genius in the person of Sam Altman. But just last week it was revealed that OpenAI is built on a shaky foundation, because not only is its position so weak financially that billions will have to be raised again in the capital market in a few months, but OpenAI's revenue and any future profits are mostly disappearing into the pockets of a happily watching Microsoft.

Musk and Altman are opposites. Altman does everything in his power to be taken seriously by the establishment; just ask his own ChatGPT anything about Altman and you will be inundated with an anthology that is hardly based on fact. According to ChatGPT, Altman isn't walking on water just yet, but it doesn't make much of a difference, and the fact that he hasn't succeeded yet is mostly due to the water.

Musk has been doing everything lately to make himself look ridiculous; from randomly giving away a million dollars a day to anyone who signs his petition and spouting nonsensical conspiracy theories, to supposedly autonomous robots pour ing cocktails. See what Marques Brownlee observed of the Tesla robots and burned down in gorty fashion.

It's the reason I deliberately didn't report anything about the robots last week, because clearly this silly gimmick served to get more publicity for Tesla. All goofiness from Musk that distracts from actually special achievements of his companies. And those accomplishments are plentiful.

A giant device caught between two Mikado sticks.
Source: BBC

Reusable Starship 

Because what almost gets snowed under is that although there are now other players in space exploration such as Blue Origin and Rocket Lab, nothing comes close to what SpaceX has achieved in terms of reusable space technology. The BBC explains in this video why Starship rocket reuse is so important.

The bottom line is that reuse greatly reduces the cost of space travel, but developing reusable rockets in particular is extremely complex. That's why no other party has yet succeeded. On images, the SpaceX Starship does not look so imposing, but the rocket weighed 3.5 million pounds when it landed, eight times as much as a full Boeing 747. The Starship is also all but six feet as tall as such a 747 is long. In short, a huge behemoth was caught between two Mikado sticks on the very first attempt.

It therefore makes sense that SpaceX was awarded an eight-launch contract worth $733 million by the U.S. Space Force on Friday, as part of a program designed to promote competition among launch providers.

Super data center in nineteen days

Then another mega-success for Musk that went underexposed. If anyone knows about AI and data centers it is Jensen Huang, CEO of Nvidia, which became worth $3.4 trillion on Tuesday. It is second only to Apple 's ($3.57 trillion) world's most valuable company and, I keep repeating, it is a matter of months, not years, until Nvidia becomes the world's most valuable company. Last week, Huang said this about what Musk's team managed in nineteen days:

"And first of all, recognition of achievement where it's deserved. From the moment of concept to a data center ready for NVIDIA to have our equipment there, to the moment we turned it on, had everything plugged in and it did its first workout. So that first part, just building a huge plant, liquid-cooled, powered, in the short time it was done, I mean, that's like superhuman. 

Yes, and as far as I know, there is only one person in the world who could do that. Elon is unique in his understanding of engineering, construction and large systems, and mobilizing resources. It's incredible. Of course, his team of engineers is extraordinary. The software team is great. The networking team is great. The infrastructure team is great. 

Elon understands this deeply. And from the moment we decided to start, the planning with our engineering team, our networking team, our infrastructure and computing team, the software team, all the prep work, then all the infrastructure, all the logistics and the amount of technology and equipment that arrived that day, and NVIDIA's infrastructure and computing infrastructure and all that technology, to the training: in 19 days."

Huang is going full on the organ here about a big customer, let that be clear. Yet we should also put that in perspective: the total order value of what Musk's Companies X and Tesla have ordered from Nvidia is estimated to be between $4 billion and $5 billion. A huge sum, but Nvidia turned over an average of $326 million in sales per day in the last quarter, so it's about roughly ten days to two weeks of sales. There are larger customers that Huang does not extol so much, and his respect for Musk is heartfelt.

Climate and health care need Musk

Starship and the new data center are two examples of Musk's unique talents. I have written it many times: precisely because Musk is so uniquely able to get huge projects successful that combine mechanical engineering with software development, it would be great if he used his talents and his ability to combat global warming.

Developing technology to remove CO2 from the atmosphere is complex and expensive; two adjectives Musk knows how to use. Instead of spending his talents preparing another planet, his attention is desired to save our current one. We'll take those silly tweets at face value.

Four truths about OpenAI

Against these successes of Musk, the situation of OpenAI pales. Earlier this month, OpenAI's insatiable hunger for capital was already covered (see: $10 billion for OpenAI), but on Thursday, the New York Times also came out with a report that things are rumbling on all sides in the relationship between OpenAI and Microsoft:

" The close partnership between Microsoft and OpenAI is showing signs of wearing out. The "best bromance in the tech world" has undergone a reality check as OpenAI has sought to change its agreement with Microsoft and the software maker seeks to reduce its dependence on the start-up."

Alex Kantrowitz saw cause for further analysis in the article and came to four troubling conclusions about OpenAI:

  1. Unpredictable training costs: OpenAI emphasized that training costs are not fixed and can be adjusted in the future. The company spends about $3 billion a year on training but excludes these costs when presenting profitability, which raises questions. The way OpenAI's newer models work may shift some of the traditional training burden to inference, making future training costs difficult to predict. Despite possible cost-saving techniques, OpenAI's focus on developing larger models will likely involve high training costs.
  2. ChatGPT as a revenue source: Contrary to expectations that OpenAI's API would generate the most revenue, the company expects ChatGPT to be the main revenue source until at least 2029. OpenAI is betting on the continued growth of conversational AI and offers different versions of ChatGPT for different markets, such as business and education. However, ChatGPT's long-term success depends on how much user interest in AI interactions increases.
  3. Significant payments to Microsoft: OpenAI is already paying Microsoft large amounts from revenue, despite not yet making a profit. A projected $700 million payout to Microsoft in 2024 highlights the financial burden of the partnership, on top of the already high costs of AI development.
  4. Future funding needs: Despite raising $6.6 billion, projected losses of $5 billion in 2024, and possibly $14 billion in 2026, indicate that OpenAI will likely need to raise capital again soon. This raises questions about the sustainability of the business model and whether investors will continue to be willing to support such high expenditures.
Geoffrey Hinton is one of the leading thinkers and critics of AI.
Source: Nobel Prizes.

Nobel laureates Hinton and Hassabis vs. OpenAI

It will happen to you that your biggest competitor and your biggest critic win a Nobel Prize in the same year. It happened to Sam Altman of OpenAI when Demis Hassabis and George Hinton both won a Nobel Prize, Hassabis for chemistry and Hinton for physics. In an interview with the Nobel Prize website, Hinton made a comparison between the fight against global warming and the regulation of AI.

"I think it's quite different from climate change. With climate change, everyone knows what needs to be done. We have to stop burning carbon. It's just a matter of the political will to do that. And big companies that make big profits don't want to do that. But it's clear what needs to be done.

Here (at AI, MF) we are dealing with something where we have much less of an idea of what is going to happen and what to do about it. I wish I had a simple solution, that if you do this, everything will be fine. But I don't have that. Especially with regard to the existential threat of these things getting out of control and taking control, I think we're at a kind of turning point in history, where in the next few years we have to figure out if there's a way to deal with that threat.

I think it's very important now for people to work on the issue of how we maintain control. We need to do a lot of research on that. I think one of the things governments can do is force the big companies to put a lot more of their resources into security research. So that, for example, companies like OpenAI can't just put security research on the back burner."

Hassabis was founder of DeepMind (sold to Google)
Source: Nobel Prizes.

Hinton had also lashed out at OpenAI last week, when he said he was proud that his former student Ilya Sutskever had fired Sam Altman as CEO of OpenAI last year. A short-lived victory, by the way, because a few days later Altman returned and Sutskever has since left OpenAI himself and started a new startup, for which he has already raised a billion dollars within three months on a $5 billion valuation.

In addition to Hinton's lashing out and Kantrowitz's rather deadly analysis, Altman was joined by news that former OpenAI CTO Mira Murati appears to be working on her own startup, for which she is sourcing talent from OpenAI.

Interpreting innovation remains difficult

Analyzing all recent developments, it appears that the much-hyped OpenAI is in increasingly difficult waters. The friction with Microsoft makes the headlines almost daily. Meanwhile, amid all his pottering, Elon Musk is building infrastructure he can enjoy for years to come, both at SpaceX with the reusable Starship and with X and Tesla in the new super data center. I don't bet, but if I did, I would never bet against Elon Musk. 

Thanks for the interest and see you on Sunday!

Categories
invest technology

Elon Musk steps into self-driving Tesla, but investors back out

Elon Musk on Thursday unfolded his vision for robotic cabs and self-driving cars. Investors reacted with disappointment on Friday, and Tesla shares fell as much as nine percent. Whatever else you may think of him; Musk's long-term vision of urban mobility is becoming increasingly relevant. With urbanization, declining car ownership and working from home, a future of shared transportation, preferably on demand, is coming ever closer.

You don't buy the Tesla Cybercab for the decorative rims.
Source: Tesla video
Singapore relentlessly carless
Last week I was in Singapore for Tech Week, one of the world's most futuristic cities, where Electronic Road Pricing (bill driving) has been working just fine for years and just the permit to buy a car costs a minimum of seventy thousand Euros.

Singapore understands that a city-state of six million people simply cannot accommodate much car traffic and acts accordingly. Only 14% of the population owns their own car.

Due to various taxes and duties, the price of a small Toyota easily exceeds two tons. Old cars are non-existent, as the permit is valid for only ten years. It is hell for a car enthusiast. Public transportation in Singapore, on the other hand, is fantastic: spotless, spotlessly clean and extremely safe, with air-conditioned buses and subways that run every three (!) minutes during rush hour.

The understandable lament of a Singaporean cab driver, who sometimes pays up to a hundred dollars a day for the rental of his cab on a turnover of a hundred and fifty dollars, made me extra fascinated by the picture of the future that Musk outlined Thursday at the Tesla event "We, Robot" at Warner Brothers studios in Los Angeles. The Verge produced a six-minute recap.

Cybercab and Roboofen
The highlight was the unveiling of the Tesla Cybercab, a fully autonomous robotic cab with no steering wheel or pedals, because there is no driver: everyone is a passenger. The Cybercab is positioned as an affordable solution under $25,000, but even the most inveterate Tesla fanboy doesn't believe that price will work.
Musk introduced the Robovan with the enthusiasm with which someone introduces their new wife to their ex-mother-in-law.
Source: Tesla video

In addition to the Cybercab, Musk also presented plans for autonomous versions of the Tesla Model 3 and Model Y, with hopes that these vehicles will hit the market by 2026. (Musk himself said "expectation," but my ears translate that directly from Elon Musks to "hope.") Surprising was the introduction of the Tesla Robovan (pronounced, according to Musk: Robooofen), a small bus shaped like a clunky iron, which can seat 20 people.

Cars 6% in use

In Musk's ideal picture, buyers would "rent out" their Tesla Cybercab or Robovan to the community as soon as they don't use it themselves. Since people use their car less than 6% per week on average, the Sacred Cow would be able to recoup its investment the remaining 94% of the time.

This may sound strange now, but who remembers when cab drivers in Europe gleefully orated that Europeans would never drive their own cabs in their Prius and that even if it happened, we exalted Europeans would never get in the car with amateurs? The skepticism against a new form of transportation will disappear very quickly the moment the Cybercab shows up quickly, the rides are safe and the price is comparable or lower than that of an Uber.

Disappointment among investors

Despite Musk's ambitious presentation, investors were not enthusiastic. Among analysts, disappointment prevailed over the lack of concrete details about planning, expected costs and potential sales for Tesla.

This reluctance among investors reflects a broader problem facing companies like Tesla. Investors often base their expectations on quarterly results, whereas Musk has a long-term vision, possibly as much as 30 years ahead, to change the world with innovative technologies. SpaceX is another example of that approach.

Musk runs Tesla like a startup

What many investors seem unwilling to understand, however, is that Musk runs his companies like startups. This means he focuses on long-term growth and value maximization and often opts for short-term tactical improvisation.

This is normal for startups that have not yet found a complete product-market fit, with apologies for the crassness. In this, Musk's approach differs fundamentally from the traditional approaches of publicly traded companies, where predictable quarterly numbers set the company's course.

The last year on the stock market was painful for Tesla with 16% decline: not only did the stock perform 50% worse than the Nasdaq Composite index, but Uber rose 88%. Source: Google Finance

For many investors, this approach is difficult to accept, especially when faced with short-term losses or delays in product launches. Musk's long-term vision - aimed at a world in which shared autonomous vehicles dominate urban transportation - may ultimately bring about a radical transformation of the mobility sector.

Investors in Tesla should not whine, because everyone knows by now that buying a stock in an Elon Musk company is more like a roller coaster ride than a ride in a gently buzzing Tesla.

2050: global urbanization

Musk's vision of robot cabs should be seen within the broader framework of urbanization, a trend that will only increase in the coming decades. By 2050, it is estimated that nearly 90 percent of the population in the US, over 80 percent in Europe and nearly two-thirds of the population in Asia will live in urban areas.

This urbanization is driven by the growth of megacities, where the demand for efficient shared mobility solutions such as robot cabs and self-driving vehicles will rapidly increase. The current model of private car ownership is becoming less and less sustainable and realistic as cities become more densely populated. This is where Musk's vision of robot cabs and shared mobility can play an important role.

From parking lot to park

By introducing autonomous vehicles, cities can not only relieve their traffic infrastructure, but also create new opportunities for more efficient use of urban spaces.

One notable aspect of Musk's presentation was his vision for the reuse of urban spaces. He suggested that parking lots in cities could be turned into recreational areas because robot cabs do not need long-term parking spaces: after all, they drive continuously.

Instead, these vehicles could be summoned via apps, similar to current services such as Uber, Lyft and Grab, but without human drivers.

Car ownership in large cities declines

Car ownership in large cities continues to decline, partly due to high costs and available public transportation. In cities such as Amsterdam, only 22% of households own a car, compared with 27% in Paris.

Even in cities such as New York and London, where car ownership is higher (around 53% and 56%), private vehicles are becoming less and less necessary due to the presence of good public transportation and alternative mobility options such as ride-sharing. However, in cities such as Los Angeles, where car ownership is still around 94%, awareness of the disadvantages of private car ownership is also beginning to penetrate, especially because of traffic congestion and the high cost of parking.

American cities, unlike Singapore, Seoul and Tokyo, for example, lack good public transportation, and it is precisely then that shared car ownership, or transportation on demand as Musk suggests, could catch on even faster.

Unexpected impact of working from home 

Another important aspect contributing to the change in urban mobility is the rise of working from home. Since the pandemic, working from home has increased by a factor of five and is now a regular part of daily life for millions of workers worldwide.

This has not only led to a decrease in commuting (and thus traffic congestion), but has also impacted the labor market, especially for groups that previously struggled to find work, such as women and people with disabilities.

The number of disabled workers in the US has increased by two million since the pandemic, largely due to the ability to work remotely. The percentage of working women in the US has also increased faster than that of men, attributed in part to the greater flexibility that working from home offers in combining work and care responsibilities. These figures provide evidence that women are both working and raising children, which makes one wonder what men are doing.

Home work sometimes horizontal work

One surprising consequence of working from home, by the way, is its potential impact on birth rates. In countries in East Asia, where long working hours and intense pressure in parenting have led to declining birth rates, working from home seems to offer parents more flexibility in combining work and parenting.

This could lead to a slight increase in birth rates, with preliminary analyses suggesting that couples who work from home at least one day a week may want an additional 0.3 to 0.5 children. In countries such as Singapore, with aging populations and low birth rates, a higher number of children per family is encouraged by the government.

Shared mobility is the future

Given increasing urbanization, the rise of working from home and the rising cost of car ownership, shared mobility is becoming increasingly logical. In cities where car ownership is declining and the use of space is changing, a model of shared cars or cars on demand offers a sustainable alternative to traditional car ownership.

Of the 10 words Elon Musk utters these days, or pours out on X over us, 11 should be taken with a hefty grain of salt. But Musk's vision of robotic cabs fits seamlessly with global developments that governments will only encourage.

The success of Airbnb suggests that enough people have little problem renting out their property at the right price. There is no reason to believe this will be any different with cars - self-driving or otherwise.

Totally useless hinged doors (try this on an Amsterdam canal), but it looks nice.

Musk is likely to play a role as a forerunner with his self-driving cars, which will then be overtaken in the mass market by other players. Just as is happening now with Tesla, which faces competition from Chinese manufacturers and the traditional automakers. Waymo and Cruise are just the first. Perhaps a cause for concern for Tesla investors, but good news for everyone else.

Categories
technology

Memecoins beat Big Tech, AI and S&P 500 by a wide margin

Big Tech in the last month: up 8% versus 4.5% for the S&P 500.

Over the past 30 days, big tech stocks Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla have averaged 8% growth. These companies remain an attractive choice for investors due to their market dominance and continued growth potential in several growing sectors, including cloud computing, e-commerce, and social media.

To put this performance in perspective, the broader market, represented by the S&P 500, also performed solidly, rising 4.5% over the same period.

This growth indicates that overall market sentiment remains positive, supported by macroeconomic factors such as falling inflation and a more stable interest rate outlook. In summary: Big Tech performed almost twice as well as the S&P 500 last month: 8% versus 4.5%.

Top 9 AI Shares: robust growth


Looking beyond traditional technology companies, AI stocks remain the star of this investment year. The top 9 AI stocks according to the AI Spotlight 9 compiled by me entirely subjectively, have shown an impressive 136% rise so far this calendar year. In the past month, these AI stocks again rose much harder than Big Tech stocks (8%) and the S&P 500 (4.5%), averaging 14%. So investors are undiminished by the potential of the big AI players.

AI Spotlight 9 update: Crowdstrike replaced by TSMC. Increase of 14% over the past month.

There is a great memecoin called 'dogwifhat'


The unexplained memecoins, often fueled by social media hype, especially on Telegram, are also experiencing heyday. Dogecoin, the largest memecoin measured by market value, has a market value of $16 billion at the time of writing (Saturday night, Oct. 5) and a daily trading volume of $430 million.

This is why smart cousins and other daredevils are in memecoins: the biggest nine memecoins rose 43% in the last month.

On average, the top 9 memecoins have experienced 43% growth over the past 30 days. This highlights that speculative interest in this market is tremendously strong, despite all the warnings from well-intentioned old men like me about volatility and lack of underlying value.


Conclusion

The past month has shown strong returns for various types of investments, from traditional big tech stocks to innovative AI companies and even memecoins. Yet it still feels like comparing Warren Buffett's performance in the stock market to your grandmother's performance at bingo.

Big tech remains a stable choice with steady growth, AI stocks show their huge growth potential, and memecoins remain a speculative but lucrative short-term option.

Categories
AI technology

Meta's glasses hacked into Terminator glasses

Sam Altman's Messiah behavior about AI is reminiscent of the way Mark Zuckerberg propagated the blessings of social media for society, before the negative effects became apparent and he was publicly keelhauled, including before the U.S. Congress. Zuckerberg has since argued that social media is a mirror for society rather than the enabler of evil, making him increasingly confident in his presentation to the outside world.

"Inside the bro-ification of Mark Zuckerberg" is the untranslatable headline above an excellent Washington Post article about Zuckerberg reinventing himself as a muscular version of ... Julius Caesar? Zuckerberg has switched from his familiar hoodie to t-shirts with self-invented lyrics quoting the Latin emperor, but apparently forgetting how Caesar met his end.

Zuckerberg is having a banner year: strong financial performance sees Meta shares rise sharply after a disastrous 2023, Apple's VR glasses flop while the new Ray-Ban Meta smart glasses get rave reviews. But once again, Zuckerberg seems blind to the problems his company's introduced technology could cause.

'Man sees unknown woman - including residential address.'

Creepiest glasses ever

Harvard students turned Zuckerberg's Meta glasses into "Terminator glasses" that allow them to do very creepy things simply by looking at someone:

"The technology, which combines Meta's smart Ray-Ban glasses with the facial recognition service Pimeyes and some other tools, allows someone to automatically go from a face to a name, phone number and home address."

The video published by the students on Tuesday paints a terrifying picture of a world in which, for example, any creep wearing such glasses could stalk any woman. The creators have shared all the information about this "I-XRAY" here.

Categories
AI technology

$10 billion for OpenAI: $6.6 vc funding and $4 billion from banks

The announcement of a major funding round in the tech world is often accompanied by a clichéd photo of a group of men in light blue shirts and the occasional rebel in a black t-shirt, trying to look tough into the camera. OpenAI, in announcing its new funding round of no less than $6.6 billion, out of a $157 billion valuation, posted an abstract image without a source; probably because there is no management left except CEO Sam Altman himself.

The text was also dry by Altman standards: "This funding allows us to strengthen our leadership in AI research, increase our computing power and build tools that help people solve complex problems."

"The dawn, blah blah blah chocolate custard"

This is less prosaic than the Bouquet Series lyrics Altman recently posted on his own blog:

"We must act wisely, but with conviction. The dawn of the Intelligence Age is a profound development with very complex and extremely risky challenges. It will not be an entirely positive story, but the benefits are so enormous that we owe it to ourselves, and to the future, to figure out how to navigate the risks that lie ahead."

This almost seems like one from the oeuvre of the unsurpassed Kees van Kooten: "but enough about myself, what do you think of my hair?" Because here Altman is unabashedly trying to attach half of his own company name to a term that is supposed to mark the period after our current post-industrial era. It is, of course, rather the dawn of the TikTok era, but TikTok is not as good at PR.

OpenAI now bronze, after SpaceX and ByteDance (TikTok)

The funding round increases the pressure on Altman to live up to this high price tag, normally via an IPO. OpenAI is now worth far more than any other venture-backed company ever before at the time of their IPO, including Meta, Uber, Rivian, and Coinbase, according to a PitchBook analysis.

Funding for OpenAI only just fits the chart. Source: Pitchbook.

OpenAI is now the third most valuable venture-backed company in the world. Only Elon Musk's SpaceX, valued at $180 billion, and ByteDance, the owner of TikTok valued at $220 billion, are worth more.

Is there any competition for OpenAI?

To put it in Dutch perspective: OpenAI is worth five times as much as Philips, as much as Unilever and still only 20% less than Shell. OpenAI's competitors must look at the new funding with horror. xAI, founded by Musk, raised more than $6 billion earlier this year, but with a valuation of "only" $24 billion. OpenAI's nearest competitor, Anthropic, was valued at $19.35 billion as recently as January.

A few days ago there was brief enthusiasm in the market when Nvidia, nota bene the party that is going to earn the most from OpenAI's new funding because it supplies all the hardware under the hood of ChatGPT, introduced a proprietary Large Language Model that seemed open source, called NVLM. Unfortunately, the technology is not allowed to be used in commercial applications, making Google with Gemini the only real remaining competitor to OpenAI. Or Anthropic needs to find funders who dare to throw billions at it.

$4 billion credit facility for OpenAI

Early investors in OpenAI are obviously in jubilant spirits, including the legendary Vinod Khosla, who in this interview with Bloomberg almost even seemed to be caught smiling. Khosla: "It is the most important tool we have ever had in human history to create abundance and realize a fairer, more equitable, prosperous society."

Khosla also reported that he has used ChatGPT for more mundane tasks - from quickly learning complex material to designing his garden. The latter is especially impressive, as Khosla's garden has a mile-long beach that has been the subject of lawsuits for years.

On Thursday, it emerged that in addition to its $6.6 billion investment round, OpenAI also managed to secure a $4 billion credit facility from a number of banks. With over $10 billion, the company can go on for a while, although the question remains how far this pole reaches for the company that loses millions every day.

Mozart of mathematics

Anyone who throws around bombastic cries about AI being the savior of humanity brings criticism upon themselves. These range from substantive criticism of the quality of the technology developed by OpenAI, to the way Altman runs the company.

More and more is leaking out about how OpenAI is rushing, especially under pressure from competition from Google, to release products that are far from ready and insufficiently tested. Even before Mira Murati's unexpected departure from OpenAI, staff complained that the o1 model was released too early.

The Atlantic advocates putting aside Altman's turgid rhetoric and looking primarily at what OpenAI's products are currently capable of:

"Altman insists that the deep-learning technology that powers ChatGPT can in principle solve any problem, at any scale, as long as it has sufficient energy, computing power and data. However, many computer scientists are skeptical of this claim and argue that several more major scientific breakthroughs are needed before we reach artificial general intelligence."

Terence Tao, a mathematics professor at UCLA, is "a real-life superintelligence, the Mozart of Mathematics," according to The Atlantic. Tao has won numerous awards, including the equivalent of a Nobel Prize in mathematics, and analyzed the performance of the OpenAI-hyped o1.

Tao's conclusions are not positive for o1's math ability and even led Tao to apologize for comparing o1's performance to that of a doctoral student.

A broader analysis appeared In the book "AI Snake Oil: What Artificial Intelligence Can Do, What It Can't, and How You Can Recognize the Difference. "In it, the authors explain why organizations are falling into the trap of AI snake oil (worthless solutions), why AI can't fix social media, why AI is not an existential threat, and why we should be much more concerned about what people will do with AI than about what AI would do on its own. I haven't read the book but the reviews are positive.

Investors lining up for AI companies

Despite all the criticism, investors are eager to invest billions in AI companies.

OpenAI has raised almost as much funding, as the competition combined. 

Reuters made this overview of funding in 2024. Adding up to over $10 billion, it especially underscores how well-funded OpenAI is, which has such an amount at its disposal due to the investment round and the banks' own credit facility.