There’s a saying in the modern world of artificial intelligence (AI): “If something isn’t possible with today’s most advanced generative AI systems, wait. It will be next month.” This maxim serves as a powerful reminder not to underestimate the capabilities of a system that is rapidly emerging and is making its way into the halls of national power, as exemplified by U.S. strategy. technology containment In great power competition with China. The competition is unfolding amid increasing participation from some Middle Eastern countries, particularly the United Arab Emirates (UAE). active role In the development of emerging technologies like AI, it is intertwined with the modern maxim of AI.
Nothing exemplifies this optimism more than reports that OpenAI CEO Sam Altman is in talks with, among other things, the UAE government to raise funding. $5 trillion to $7 trillion (Yes, trillion) if increased chip building abilityfollowing the news Altman sought. billions of dollars For chip companies focused on Tensor Processing Units.Newly established Abu Dhabi-based investment fund MGXChaired by Sheikh Tahnoun bin Zayed Al Nahyan, UAE National Security Adviserearly” is in talks with OpenAI to help fund Altman’s chip-building efforts.
The fact that such initiatives are being considered shows that the modern maxim of AI is gaining ground among a diverse range of actors eager to leverage it. Generative AI models will only continue to become more sophisticated, and nations seeking to play a leading role in the future global economy will be irreversibly disadvantaged in the long term if they do not act decisively now. There is a recognition that this will happen.
It’s no wonder that generative AI is already realizing the transformative vision the AI field has been striving for for years. However, it is questionable whether these models will deliver on the promise of innovative applications, as commercial and technical obstacles are underestimated. It echoes Matt O’Shaughnessy’s warning when speaking about “.super intelligence‘ will lead policymakers astray, and Altman’s strategically ambitious but arrogant approach threatens to lock AI policymaking into a future that may never materialize, prompting policymakers in the U.S. and the Middle East alike to You should be wary.
as financial times According to reports, the (anonymous) chief financial officer of a multi-billion dollar company cited the disparity in peer spending on products such as Amazon Web Services (AWS) and ChatGPT, noting that the latter “often drives business They were maxed out and he was surprised at how little they were actually paying for OpenAI.”
moreover, Yoon Jun They argue that the market for chips that power AI is starting to show signs of overheating. Coupled with the lack of widespread adoption by companies, generative AI is very similar to the boom and bust of telecom stocks during the dot-com bubble.
Furthermore, according to the reporter: information shows that commercial adoption of generative AI is slower than the hype due to a mismatch between expected functionality and current costs. Customers of cloud providers such as Microsoft, AWS, and Google are reluctant to spend on new AI services given the high cost of running the software, the shortcomings in accuracy, and the difficulty in determining the amount. They are cautious or intentional about increasing their numbers. the value they get from it. ” In private, short term expectations Returns on investment in AI implementation were suppressed.
To be sure, generative AI companies have short-term advantages.altman is impressive strategic actormoving rapidly to capitalize on the unexpected success of ChatGPT-3.5 in late 2022, and deftly adjusting his rhetoric about the future. technology And the role of OpenAI in this is us senate and international audience. OpenAI further 2 billion dollars Within just 10 years of its inception, annual revenue has increased (up from $1.3 billion in mid-October 2023).
In addition, Salesforce’sEinstein GPTGenerative AI customer relationship management (CRM) technology and Swedish FinTech companies Khurana’s Handle customer service chats using GPT-4.
Finally, OpenAI’s text-to-video generator “Sora” is the culprit. anxiety inside entertainment industry.this is on the back Hollywood writers’ 148-day strike In 2023, AI-generated materials took center stage. OpenAI is currently trying Convince Hollywood movie studios and directors with Sora.
Nevertheless, generative AI systems are being deployed very rapidly. in front A strange chain of events occurs in which their appropriate uses are identified.as alan holland Venture capital investments in generative AI systems were made by an absurd reversal of basic investment practices: first identifying the business problem; after that Build an application to solve it.
The technical hurdles facing generative AI are equally severe. Hallucinations are, for example, innate characteristics The percentage of technology that can be reduced but not eliminated in real-world situations.
Furthermore, large-scale language models (LLMs) include plan and reason. Led by AI research group Subbarao Kambanpatti I discovered that LLMs engage in what they say, rather than planning or reasoning in a human-like way. Approximate search—They complete given input by reconstructing its completion word by word. However, in contrast to traditional databases, novelty The output ratio of LLM means that the process of approximate search reconstructs the data rather than retrieving the exact data found. probabilistically.their the study LLMs, including GPT-3.5 and GPT-4, have shown clearly poorer ability to execute plans when tested autonomously. This is an important condition that allows LLM to improve the search process. external Planning systems through idea generation.
Finally, generative AI systems lack: intellectual autonomy. Simply put, the basic conceptual work is still done by humans. Humans decide which concepts these systems exploit and judge whether the output is logically consistent. It is not clear how his GPT-4 descendants, which belong to the same basic architecture, can alleviate this problem.
Considering these statements, modern maxims about AI should be justified. When a generative AI system appears to have indisputably crossed another threshold, just wait. The end result will be walkbacks and warnings. go back to reality.
State and private stakeholders, including Abu Dhabi, should pay close attention to Mr. Altman’s proposals. It determines the underlying architecture of the AI system. range We will discuss its capabilities and the actual costs of its construction, training, and performance. AI infrastructure is inseparable linked to the architecture of the system. Generative AI is no exception. Therefore, any attempt to plan the infrastructure needed for future AI necessarily involves a degree of uncertainty, as future AI systems may employ unfamiliar architectures. It will be.
nothing new identify The future of AI training and performance will likely leverage different hardware. Identifying that the paradigms in which such decisions are made may be potentially unknown to today’s researchers is less appreciated. Grady Booch is a software engineer. unified modeling language (UML)—ha vocal Regarding the foolishness of Altman’s proposal, Identify The broader trend is “a clear sign that modern AI has chosen the wrong architecture.”
altman teeth He’s sending out an invitation to the future with his $7 trillion idea, but this technology may have another future entirely unknown to us.
Vincent J. Kalkidi He is a non-resident scholar in the Strategic Technology and Cyber Security Program at the Middle East Institute. He is also a member of his 2024 Next Generation Cohort in American Foreign Policy. His opinions are his own.you can follow him linkedin or X.
Copyright 2024 Nexstar Media Inc. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.
