Curious AI 36
Oliver Rochford
Evangelist @ Auguria | Technologist | Cyberfuturist | Startup Advisor | Former Gartner Analyst
Welcome to issue 36 of the Curious AI Newsletter, curated by Oliver Rochford , Cyber futurist and former Gartner Research Director, and synthesized and summarized using AI.
AI Tribe of the Week
AI-Skeptic
Doubts the hype around AI, always questioning whether it will actually deliver on its promises or just create more problems.The party poopers of the AI world. “Sure, AI can beat humans at chess, but can it cook a perfect steak? Didn’t think so.”
Tagline:“AI? More like ‘Eh, why?’”
Want to discuss AI, quantum, and other emerging technologies?
Join our Slack!
I am speaking at the Emerging Tech Community Roundtable
I am thrilled and grateful that The Jobs of The Future has invited me to speak at their Emerging Tech Community Roundtable, alongside Ali Hussein Kassim and Dr. George Tumanishvili (PhD).
We will be discussing how AI will affect the workplace and business, moderated by Anastasia "Tracy" Raissis and Xiaochen Z.
I am really looking forward to this, and I would love it if you joined us:
When: August 6th at 10AM EST | 9AM CST.
RSPV below:https://lnkd.in/ehkQctc9
Most Batshit Crazy: No One Is Ready for Digital Immortality
"Boy, I'm like them huge fuckin' lizards, you know? Had themself two goddam brains, one in the head an' one by the tailbone, kept the hind legs movin'. Hit that black stuff and ol' tailbrain jus' kept right on keepin' on."
The Dixie Flatline to Case in Neuromancer
I have often highlighted articles about the worrying trend of making digital "replicas" of people who have died. AI technologies that promise digital immortality highlight a troubling trend: the oversimplification of human personality and, ultimately, identity.
Using fragmented data and snapshots from social media to recreate a person is not at all enough to model the depth of human experience. So many parts of our personal lives and subtleties of who we are can not be captured digitally. Think about all the things you think but never say out loud or share them with other people. From a technological point of view, it is also silly. We are not even close to being able to upload yourself, despite what some old futurists who are thinking about their own deaths might think.
It’s really a form of technological mimicry, leading to a shallow and incomplete representation of these individuals. AI will imitate anything that was not directly captured about a subject based on generic training data. This is similar to how AI can be used to make an image bigger than its original frame. It will make guesses. Critics are right to say that this kind of technology will hurt more than help because it will show a distorted version of loved ones that might make you feel better in the short term but never really connects you with them.
The business side of digital immortality also makes it possible for people to take advantage of the memories and identities of the dead, changing them or selling them without respecting their true legacy.
This means that strong legal frameworks and moral guidelines are needed to protect people's digital afterlives from becoming just products in the tech market, even if it is not technically possible to authentically copy a user.
What kind of experience would you have with this kind of technology if we made it? Would it be how William Gibson described his “Dixie Flatline" construct in Neuromancer, a digital consciousness that, while functional, lacked the true essence of the person it was modeled after? When I die, please leave me in meatspace.
"Do me a favor, boy. This scam of yours, when it's over, you erase this god-damned thing."
The Dixie Flatline to Case in Neuromancer
Most Underwhelming: OpenAI Launches SearchGPT to Rival Google
OpenAI launched SearchGPT last week. SearchGPT wants to compete with Google's search dominance by giving users direct, real-time answers and visual content instead of links. If the idea does not sound very new, that is because it isn't. Perplexity was the first to try this idea, but they had a lot of problems with it, like accusations of plagiarism. The current king of search, Google, also thought about the idea, but made a u-turn after issues with giving wrong answers and hallucinations. That is likely the reason why SearchGPT is currently only open to a small group of people.
OpenAI is expected to lose up to $5 billion this year, so search, or more accurately, advertising, probably seems like one of the few things that can make the company using GenAI any real money in the near future. The future of AI is apparently in advertising.
Most Ironic: Using the Term 'Artificial Intelligence' in Product Descriptions Reduces Purchase Intentions
A study by researchers at Washington State University found that mentioning "artificial intelligence" in product descriptions decreases purchase intentions, particularly for high-risk products like expensive electronics or medical devices. This counterintuitive outcome suggests marketers should avoid highlighting AI gratuitously to maintain consumer trust and increase the likelihood of a purchase.
Intrigued by the most recent developments in quantum technology?
Check out our Intriguing Quantum Newsletter
Sovereign AI & AI Nationalism
Europe Launches 'AI Factories' Initiative
Europe's supercomputing body, EuroHPC JU, is launching AI Factories to bolster the region's AI capabilities. These AI Factories will provide compute power, data, and talent, creating hubs for AI advancements in health, energy, manufacturing, and more. The initiative aims to support startups, SMEs, and academia, with funding estimated at €400 million for 2024 and up to €800 million by 2027.
Sentiment: Positive | Time to Impact: Mid-term
Spain Poised to Become Europe’s Data Center Capital
Spain is set to become Europe's leading data center hub, attracting major investments from Amazon and Meta. Amazon plans to invest €15.7 billion in Aragon, while Meta is building a data center near Talavera de la Reina. These investments aim to decentralize data centers from Madrid to less populated areas.
Sentiment: Positive | Time to Impact: Mid-term
AI Copyright, Regulation, and Antitrust
AI Existential Risk Probabilities Are Too Unreliable to Inform Policy
The article argues that AI existential risk probabilities are too speculative to inform policy. It critiques the methods used to estimate these risks, highlighting their unreliability due to the lack of a reference class, the absence of robust theoretical models, and the variability of subjective probabilities. The authors suggest that such estimates are more reflective of individual biases than credible risk assessments and warn against using these unreliable probabilities for policymaking.
Sentiment: Negative | Time to Impact: Short-term
AI Business
Meta's $50 Billion VR Gamble Fails to Pay Off
Meta’s Reality Labs division has lost nearly $50 billion over the past decade, primarily due to investments in virtual reality (VR), augmented reality (AR), and the metaverse. Despite substantial financial outlays, these ventures have yet to gain significant traction. CEO Mark Zuckerberg remains optimistic, viewing these losses as long-term investments.
Sentiment: Negative | Time to Impact: Long-term
Zuckerberg’s Vision of the Future: A Poorly Rendered AI Shrimp In Every Social Feed
Mark Zuckerberg's latest AI vision involves AI-generated content flooding social media feeds, including poorly rendered avatars mimicking influencers. This approach has raised concerns about the quality and authenticity of online interactions, with critics questioning the value and potential backlash of such AI-driven content proliferation.
Sentiment: Negative | Time to Impact: Short-term
A CIO Canceled a Microsoft AI Deal. That Should Worry Tech Industry.
A pharmaceutical company’s CIO canceled a deal for Microsoft’s Office 365 Copilot AI, citing high costs and low value. The AI tool, used by 500 employees, was likened to “middle school presentations.” The decision raises concerns about the practical value and cost-effectiveness of AI tools in enterprise settings, potentially impacting tech industry adoption and investment.
Sentiment: Negative | Time to Impact: Immediate
AMD Q2 2024 Earnings: AMD’s AI GPU sales just went from a billion dollars cumulatively to a billion dollars quarterly.
AMD reported strong Q2 2024 results, with revenues of $5.8 billion, surpassing expectations. Key drivers were record data center revenue of $2.8 billion and robust AI growth, primarily from MI300X AI chips. Despite declines in gaming and embedded segments, AMD's AI and data center segments position it for continued growth.
Sentiment: Positive | Time to Impact: Short-term
OpenAI is Talking to $700 Billion Chip Giant to Rival Nvidia
OpenAI is in discussions with Broadcom, a $700 billion chip giant, to develop AI chips that could rival Nvidia's. This move is part of OpenAI's strategy to reduce dependency on Nvidia and potentially boost its AI infrastructure capabilities. Broadcom's extensive expertise in semiconductor and infrastructure solutions positions it as a strong contender in this collaboration.
Sentiment: Positive | Time to Impact: Mid-term
Recommended by LinkedIn
OpenAI Launches SearchGPT to Rival Google
OpenAI has introduced SearchGPT, a prototype AI-powered search engine designed to challenge Google's dominance in the search market. This tool aims to provide real-time, direct information responses instead of traditional links. SearchGPT also supports visual responses, such as images and videos, and emphasizes clear source attribution to avoid plagiarism issues seen with competitors like Perplexity. Initially available to a limited group of users, the prototype seeks feedback for future improvements.
Sentiment: Positive | Time to Impact: Mid-term
Concerns Over AI Bubble Amid Big Tech Investments
Goldman Sachs has raised concerns about a potential AI bubble due to the massive investments made by tech giants like Amazon, Microsoft, Google, and Meta. These companies have collectively spent over $1 trillion on AI infrastructure, driven by the belief that AI will significantly boost productivity and GDP. However, Goldman Sachs analysts, including Jim Covello, warn that the high costs may not yield expected returns, and transformative use cases for AI have yet to emerge. This skepticism suggests that if significant AI applications do not materialize within the next 12-18 months, investor enthusiasm is likely to wane, leading to market volatility
Sentiment: Negative | Time to Impact: Short-term
What Could Kill the $1 Trillion Artificial Intelligence Boom
The article explores the potential threats to the ongoing $1 trillion AI boom. Key risks include overinvestment leading to an AI bubble, the high costs of AI infrastructure outweighing returns, and the lack of transformative AI applications. Regulatory challenges, cybersecurity threats, and public skepticism about AI’s benefits further compound these risks. If significant use cases don't emerge soon, investor confidence may falter, potentially destabilizing the market.
Sentiment: Negative | Time to Impact: Short-term
Silicon Valley’s Trillion-Dollar Leap of Faith
Tech giants are investing heavily in AI, expecting transformative benefits, but profitability remains uncertain. Analysts question whether the immense spending will yield returns, as current AI applications struggle to generate significant revenue. Despite this, fear of missing out drives continued investment, suggesting a potential AI bubble.
Sentiment: Negative | Time to Impact: Short-term
Releases and Announcements
Google's Tiny AI Model Gemma 2 2B Challenges Tech Giants in Surprising Upset
Google's new AI model, Gemma 2 2B, with just 2.6 billion parameters, is outperforming much larger models like GPT-3.5 and Llama 3. This compact model demonstrates significant efficiency and accessibility improvements, running on a variety of hardware including laptops and cloud environments. The introduction of safety and transparency tools like ShieldGemma and Gemma Scope further enhances its robustness, aiming to set new standards in the AI industry.
Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025
Gartner forecasts that 30% of generative AI projects will be abandoned by 2025 due to high costs, poor data, and unclear ROI. The financial burden and impatience for quick returns challenge the viability of these projects.
AI at Work
'Is it even worth it?': Confessions of an Agency Exec on Client Challenges with Generative AI Peak
An agency executive reveals challenges with generative AI, particularly client apprehension over "AI info" tags on content, which can undermine credibility. Despite AI's efficiency, concerns about transparency and authenticity affect its adoption. Agencies must balance efficiency gains with client comfort regarding AI use in marketing.
Sentiment: Neutral | Time to Impact: Short-term
Using the Term 'Artificial Intelligence' in Product Descriptions Reduces Purchase Intentions
A study by Washington State University found that mentioning "artificial intelligence" in product descriptions decreases purchase intentions. The term reduces emotional trust, particularly for high-risk products like expensive electronics and medical devices. Marketers are advised to avoid highlighting AI in product descriptions to maintain consumer trust and enhance purchase likelihood.
Sentiment: Negative | Time to Impact: Short-term
JPMorgan Pitches In-House Chatbot as AI-Based Research Analyst
JPMorgan Chase has rolled out LLM Suite, an AI tool modeled after OpenAI’s ChatGPT, to about 50,000 employees. Designed to assist with writing, idea generation, and document summarization, LLM Suite aims to boost productivity within the asset and wealth management division. This proprietary platform underscores JPMorgan’s strategy to leverage AI while ensuring client data remains secure.
Sentiment: Positive | Time to Impact: Short-term
AI and Society
No One Is Ready for Digital Immortality
As AI technology enables digital replicas of deceased loved ones, society faces ethical and emotional challenges. Tools like deathbots can offer comfort but also complicate grief and consent issues. The potential for misuse and the need for legal frameworks are growing concerns. End-of-life planning now includes digital legacies, raising complex questions about digital immortality.
Sentiment: Concerned | Time to Impact: Long-term
The Harm of AI: Backlash Against Artificial Intelligence and Human Labor
The article discusses the growing backlash against AI, highlighting concerns over job displacement and the devaluation of human labor. Critics argue that AI technologies are causing economic and social harm by automating tasks traditionally performed by humans, leading to job losses and increased inequality. The debate underscores the need for regulatory frameworks to manage AI's impact on society.
Sentiment: Negative | Time to Impact: Mid-term
AI Trust, Risk, and Security Management
AI Browser Extensions Are a Security Nightmare
AI-powered browser extensions pose significant security risks. These include straightforward malware, data breaches, and legal issues related to copyright and plagiarism. Users often unknowingly expose sensitive data, which can be misused or incorporated into training data. Companies are advised to educate employees, allowlist safe extensions, and maintain visibility over extension usage to mitigate risks.
Sentiment: Negative | Time to Impact: Short-term
AI Ethics
Should You Be Nice to AI Chatbots Such as ChatGPT?
The article explores the ethics and practical considerations of interacting politely with AI chatbots like ChatGPT. While these chatbots do not have feelings or consciousness, treating them with respect can foster better human-computer interaction practices and encourage users to engage more thoughtfully. Experts suggest that respectful interactions with AI can help improve the overall user experience and promote positive behaviors in other aspects of digital communication.
Sentiment: Neutral | Time to Impact: Long-term
Google Sets Behavior Standards for Gemini Chatbot and its Users
Google has implemented strict behavior standards for its Gemini chatbot to ensure it remains helpful while avoiding harmful or offensive outputs. The guidelines emphasize user safety and the prevention of misinformation. Gemini is now powered by the advanced "Gemini 1.5 Flash" model, which enhances its efficiency and quality. The new version supports 40 languages and is available in 230 countries, aiming to provide more accurate and helpful interactions.
Sentiment: Neutral | Time to Impact: Short-term
AI and Robotics
Elon Musk Said Tesla Robotaxi Skeptics Should Try 'Full Self Driving.' A Wall Street Analyst Nearly Crashed
Tesla CEO Elon Musk encourages skeptics of the company's robotaxi vision to test its Full Self-Driving (FSD) feature. However, William Stein, an analyst with Truist Securities, reported a near-crash during a test drive, citing numerous illegal maneuvers and critical interventions to prevent accidents. Stein’s experience raises doubts about the readiness and safety of Tesla’s autonomous technology, which Musk claims is pivotal for the future and valuation of Tesla.
Sentiment: Negative | Time to Impact: Mid-term
Interesting Papers & Articles on Applied AI
Faulty Nvidia H100 GPUs and HBM3 Memory Caused Half of Failures During LLaMA 3 Training
Meta's LLaMA 3 training on 16,384 Nvidia H100 GPUs experienced significant hardware issues, with a failure occurring every three hours. Half of the 419 unexpected interruptions over 54 days were due to faulty GPUs and HBM3 memory. Despite these challenges, Meta achieved over 90% effective training time through advanced diagnostics and mitigation strategies.
About the Curious AI Newsletter
AI is hype. AI is a utopia. AI is a dystopia.
These are the narratives currently being told about AI. There are mixed signals for each scenario. The truth will lie somewhere in between. This newsletter provides a curated overview of positive and negative data points to support decision-makers in forecasts and horizon scanning. The selection of news items is intended to provide a cross-section of articles from across the spectrum of AI optimists, AI realists, and AI pessimists and showcase the impact of AI across different domains and fields.
The news is curated by Oliver Rochford, Technologist, and former Gartner Research Director. AI (ChatGPT) is used in analysis and for summaries.
Want to summarize your news articles using ChatGPT? Here's the latest iteration of the prompt. The Curious AI Newsletter is brought to you by the Cyber Futurists.
Co-Founder of Altrosyn and DIrector at CDTECH | Inventor | Manufacturer
3moHitting 1500 subscribers is a fantastic achievement! The "Digital Immortality" piece is sure to spark debate about consciousness uploading and the ethical implications of simulated existence. OpenAI's SearchGPT feels like a step in the right direction, but can it truly compete with Google's entrenched search dominance? The irony of AI-laden marketing is palpable are consumers becoming desensitized to the term "artificial intelligence"? Does this signify a shift towards neuromarketing strategies that bypass conscious reasoning and tap into subconscious desires through subliminal messaging and emotional triggers?