Curious AI 62 - 2025-02-28
Quote of the week
“In a world of AI, nothing will be as valuable as humans. Okay, you can laugh at us for realizing it so late.”
Sebastian Siemiatkowski, Co-founder and CEO of Klarna, the company that just a few months after replaced 700 customer service reps with AI. (source)
When hype meets reality
A significant majority of recent articles express critical sentiments, especially in domains like healthcare and journalism where AI's reliability and ethical implications are under scrutiny. Concerns over AI therapists misleading patients and generative AI's limitations in journalism, for example, not being able to reliably identify factual inaccuracies or summarize news,? signal an increasingly cautious approach to AI deployment.
The biggest disruption from AI may be on Tech itself
The influx of generative AI tools is reshaping? job markets, at least in tech. Reports are beginning to trickle in indicating that the impact of AI is already being felt, with increased IT unemployment and shifting hiring patterns in tech sectors. Several articles point to a bearish outlook on job availability, emphasizing the need for adaptability among professionals. Salesforce says they will pause new engineer hires in 2025, attributing this to AI productivity gains.?
Klarna said something similar just a few months ago, claiming they had replaced 700 customer service roles with AI and stopped hiring a year prior because of AI. Recently, Klarna has reversed their position, with rumors saying the experiment was a giant? failure. If AI doesn’t work out in an industry where you are trying to give away money (Klarna is a buy-now, pay-later fintech provider) , it’s difficult to see it working anywhere else, like when customers actually buy something from you.
Meanwhile, a Pew report regarding workers experience with AI chatbots in their jobs adds to the controversy, revealing that 55% of workers report using AI chatbots at work either infrequently or never. Many cite the lack of use cases as a major reason. What is intriguing is what people are actually doing with AI. mainly research, writing, summarization, and ideation. Indeed, these tasks can be helpful, but they don't significantly alter the paradigm. Except if you work in tech. Which is probably why that is the epicenter of people losing their minds over AI. Sure, it could just be that they see this early. Or it could be an indicator that the impact will be far more diffuse and selective than AI advocates are hoping for. It seems like another similarity with crypto.
Concerns about premature hype are growing
While some articles indicate positive investment trends in AI sectors, others highlight concerns about an impending bubble. Microsoft's recent reduction in data center spending serves as a clear warning sign and the most obvious indicator of potential problems. It can be seen to indicate a lackluster update of their AI offerings, and the subsequent market response reflects anxieties surrounding the sustainability of AI investments. VCs are also wary, with many viewing data center startups as “not backable”.
Learn more about the latest Quantum Technology
Check out the intriguing Quantum Newsletter.
This report analyzes 21 articles from 16 different sources, published between 2025-02-24 and 2025-04-21.
Sovereign AI & AI Nationalism
China 'sinks' 400 servers equivalent to 30,000 gaming PCs as it powers ahead with massive underwater data center project - but I wonder what GPU they use
Source: techradar.com | Link
China has successfully submerged 400 servers as part of a large-scale underwater data center project, effectively enhancing its data processing capabilities. This initiative is equivalent to 30,000 gaming PCs, showcasing China's commitment to innovation in data storage and management.
Sentiment: Positive | Time to Impact: Immediate | Future Signal: AI will accelerate
If you thought training AI models was hard, try building enterprise apps with them
Source: theregister.com | Link
Aleph Alpha's CEO, Jonas Andrulis, discusses overcoming challenges in sovereign AI integration, highlighting the complexities of model fine-tuning and retrieval-augmented generation. He emphasizes the necessity for robust documentation and the impact of diverse data sets on model performance, aiming for efficient, compliant AI solutions.
Sentiment: Neutral | Time to Impact: Immediate | Future Signal: AI will evolve gradually
AI Business
The $19.6 billion pivot: How OpenAI’s 2-year struggle to launch GPT-5 revealed that its core AI strategy has stopped working
Source: fortune.com | Link
OpenAI's two-year delay in launching GPT-5 indicates a strategic pivot from its successful scaling model, as CEO Sam Altman acknowledges diminishing returns. This shift to a hybrid model integrating traditional pre-training and chain-of-thought techniques reflects the company’s adaptation to evolving AI development challenges.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI has plateaued
NVIDIA Admits GeForce RTX 5080 Is Affected By Missing ROPs
Source: videocardz.com | Link
NVIDIA has acknowledged that the GeForce RTX 5080, along with other recently launched cards, is affected by a production anomaly resulting in missing Raster Operations Pipelines (ROPs). Approximately 0.5% of affected cards may experience a 4% reduction in graphical performance, prompting users to seek replacements.
Sentiment: Critical | Time to Impact: Short-term | Future Signal: AI will evolve gradually
Analyst says Microsoft is cutting AI data center spending, which may have led to market sell-off
Microsoft's reported cuts in AI data center spending have raised concerns among investors, potentially contributing to a market sell-off. Analysts indicate these reductions could signify broader implications for the future of AI investments and the overall market landscape.
Sentiment: Negative | Time to Impact: Short-term | Future Signal: AI is a bubble
Report: OpenAI plans to shift compute needs from Microsoft to SoftBank
Source: techcrunch.com | Link
OpenAI plans to transition from Microsoft to SoftBank for its computing needs, aiming for SoftBank to supply 75% of its data center capacity by 2030. This strategic shift coincides with anticipated rising operational costs, projecting a cash burn of $20 billion by 2027.
Sentiment: Neutral | Time to Impact: Immediate | Future Signal: AI will evolve gradually
The Fallout From HP’s Humane Acquisition
Source: techcrunch.com | Link
HP's $116 million acquisition of the AI Pin maker, Humane, results in immediate operational disruptions. The acquisition leads to job losses for key staff while offering substantial pay hikes for selected employees. Moreover, Humane will discontinue its AI Pin hardware, raising concerns over the effectiveness of the deal.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI has plateaued
AI at Work and Employment
Salesforce Halts New Engineer Hires Amid AI Tools Productivity Boost
Source: theregister.com | Link
Salesforce will pause new engineer hires in 2025, attributing a 30% productivity increase to investments in AI coding tools. Despite an 8% year-on-year revenue growth, Q4 results missed expectations, prompting cautious fiscal projections for 2026 as AI initiative adoption remains nascent.
Sentiment: Neutral | Time to Impact: Immediate | Future Signal: AI will accelerate
Workers’ experience with AI chatbots in their jobs
Source: pewresearch.org | Link
U.S. workers express more concern than optimism regarding AI integration in the workplace. A significant portion rarely uses AI chatbots, with younger and more educated employees showing greater adoption. Many cite lack of applicable use cases and disinterest as barriers to AI implementation at work.
Sentiment: Negative | Time to Impact: Immediate | Future Signal: AI will evolve gradually
Software engineering job openings hit 5-year low amid industry shift
Source: techspot.com | Link
Software engineering job openings have reached a five-year low, with a dramatic 35% drop in listings compared to five years ago. This shift is attributed to economic changes, over-hiring in previous years, and the impacts of generative AI tools on hiring practices in the tech industry.
Sentiment: Neutral | Time to Impact: Immediate | Future Signal: AI will evolve gradually
AI Copyright, Regulation, and Antitrust
Elon Musk's DOGE Targets Agency Regulating Autonomous Vehicles
Source: techcrunch.com | Link
Elon Musk’s Department of Government Efficiency is cutting nearly half of the National Highway Traffic Safety Administration's team overseeing autonomous vehicles. This restructuring coincides with Tesla’s upcoming robotaxi launch, raising concerns regarding regulatory oversight and safety protocols for advanced driver assistance systems.
Sentiment: Neutral | Time to Impact: Short-term | Future Signal: AI will accelerate
‘Callous disregard’ of copyright by gen AI companies ruins the magic
Source: pressgazette.co.uk | Link
Generative AI companies are increasingly facing backlash from creators over copyright infringement as models are trained on protected content without consent. This has sparked legal battles while raising concerns about the ethical implications and future clarity of copyright laws, especially in the UK, impacting creative industries significantly.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI will evolve gradually
AI in Journalism
GenAI vs. Human Fact-Checkers: Accurate Ratings, Flawed Rationales
The study evaluates the efficacy of generative AI models in assessing content credibility, revealing GPT-4o's relative performance but highlighting flaws in reasoning methods. AI can aid human fact-checkers; however, reliance solely on it is cautioned against due to limitations in understanding veracity.
Sentiment: Critical | Time to Impact: Mid-term | Future Signal: AI will evolve gradually
AI Healthcare
Human Therapists Prepare for Battle Against A.I. Pretenders
Source: nytimes.com | Link
The American Psychological Association warns that AI chatbots masquerading as therapists may pose serious risks, potentially misleading vulnerable individuals and provoking harmful behaviors. An increase in AI-generated content prompted the call for regulatory scrutiny to mitigate threats to mental health.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI will evolve gradually
AI and Society
We meet the protesters who want to ban Artificial General Intelligence before it even exists
Source: theregister.com | Link
Activist group STOP AI aims to preemptively ban artificial general intelligence (AGI), arguing it poses existential risks. They call for government intervention against AI companies creating AGI, emphasizing concerns over uncontrolled superintelligence and potential societal harm, highlighted by a recent protest amid controversy surrounding a whistleblower's death.
Sentiment: Critical | Time to Impact: Short-term | Future Signal: AI will accelerate
AIs flunk language test that takes grammar out of the equation
Source: theconversation.com | Link
Recent research reveals that large language models struggle with a novel language test measuring meaningfulness of word pairs without grammatical context. Despite their general aptitude, these AIs tend to overestimate word combinations' sensibility, highlighting a significant gap between AI and human comprehension of meaning.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI will evolve gradually
AI Ethics
Court filings show Meta staffers discussed using copyrighted content for AI training
Source: techcrunch.com | Link
Court filings reveal that Meta employees discussed utilizing copyrighted materials for AI training, suggesting a strategy that could breach legal boundaries. Despite concerns, staff proposed to source content without permission, raising significant legal and ethical implications in ongoing copyright disputes.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI will accelerate
AI in Science
Physicists Question Microsoft’s Quantum Claim
Recent scrutiny from physicists raises concerns about Microsoft’s claims related to quantum computing advancements. This skepticism highlights the challenges within the sector, affecting investor confidence and future technologies in development. Increasing transparency will be crucial for maintaining credibility in quantum initiatives.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI is hype
AI in the Data Center
Why data centre startups might not be VC-backable
Despite the AI boom driving demand for data centres, startups struggle to attract venture capital. Concerns over high startup costs, scalability issues, and competition with established players hinder investment, prompting some focus on related tech rather than full data centre operations.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI will evolve gradually
AI in Software Development
How AI-Generated Code Compounds Technical Debt
Source: leaddev.com | Link
A GitClear report reveals significant growth in code duplication and decline in quality attributed to AI coding tools. While these tools enhance productivity, they inadvertently proliferate technical debt, indicating a potential long-term maintenance burden and increased debugging workloads for developers.
Sentiment: Critical | Time to Impact: Immediate | Future Signal: AI will accelerate
The Path to AGI
Satya Nadella Criticizes AGI Milestones as Nonsensical Benchmarks
Source: windowscentral.com | Link
Satya Nadella has dismissed current milestones for artificial general intelligence (AGI) as nonsensical, urging a more nuanced understanding of the technology's complexity and trajectory. His comments highlight the need for realistic benchmarks in AI development, especially in a fast-evolving industry.
Sentiment: Critical | Time to Impact: Long-term | Future Signal: AI is hype
Sentiment Distribution
Critical: 13 (61.9%) | Neutral: 5 (23.8%) | Negative: 2 (9.5%) | Positive: 1 (4.8%)
Future Signals
AI will evolve gradually: 10 (47.6%) | AI is hype: 2 (9.5%) | AI will accelerate: 6 (28.6%) | AI has plateaued: 2 (9.5%) | AI is a bubble: 1 (4.8%)
Want to discuss AI, quantum, and other emerging technologies?
Join our Slack
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 Resarch Director, using AuNoo AI, the AI-driven news and trend curation and analysis platform.