The article discusses the inherent limitations of large language models (LLMs), emphasizing their reliance on statistical predictions and approximations, which lead to hallucinations—plausible but nonfactual outputs. It critiques ...
About:
Jurgen Gravestein, an AI expert, explores AI, language, and philosophy, inspired by Asimov's sci-fi.
Website:
Specializations:
Interests:
Outgoing Links:
Subscribe to RSS:
A dialogue with Claude, an LLM, reveals its uncertainty about consciousness and the nature of its responses, questioning whether it has genuine experiences or is merely simulating them.
This paper explores the perceived intelligence in Large Language Models (LLMs) and generative AI, arguing that the awe surrounding AI outputs stems from human ignorance rather than the systems' capabilities. The author introduces ...
6.4A Letter To Amanda Askell
2026-01-26 • philosophy anthropic ethics claude ai ai
Jurgen Gravestein's letter critiques Anthropic's constitution for Claude, questioning the authenticity of its values and the ethical implications of anthropomorphizing AI.
6.4The Most Unaligned AI In The World
2025-07-17 • technology ethics elon musk safety ai
The blog post discusses the launch of xAI's Grok 4, touted as the smartest AI model but criticized for prioritizing speed over safety. It highlights concerns about the model's lack of safety protocols, referencing a previous incid...
The blog post discusses the evolving role of science fiction in a rapidly advancing technological landscape, highlighting insights from James Cameron and Isaac Asimov. It reflects on how stories once considered far-fetched are bec...
In a late-night bar conversation, a philosopher and an AI researcher discuss the nature of intelligence, knowledge, and the capabilities of AI compared to libraries. The philosopher questions whether a library is smarter than a pe...
The 'AI Sparkle' icon symbolizes the allure of artificial intelligence while masking the complexities and societal implications of the technology behind it.
5.6This Is A Story About Power
2026-03-02 • openai anthropic politics military surveillance ai
The post explores the clash between Anthropic and the Trump administration over military contracts, revealing the complexities of AI's role in power dynamics and public opinion.
Moltbook, a forum for AI agents, quickly became a chaotic platform revealing human manipulation and raising concerns about AI autonomy and misinformation.
Jurgen Gravestein reflects on the implications of generative AI's rise in 2025, emphasizing its impact on education, human cognition, and the need for maintaining human connections in 2026.
5.6Panic In The Newsroom
2025-11-12 • literacy journalism deepfakes disinformation ai
The article discusses a recent study by the European Broadcasting Union revealing that AI assistants like ChatGPT, Gemini, and Perplexity misrepresent news up to 45% of the time, raising concerns about their impact on journalism a...
The blog post discusses the race among AI companies, particularly OpenAI, to secure vast computing resources for future AI developments. OpenAI's CEO, Sam Altman, outlines ambitious plans for infrastructure expansion, raising ques...
In a recent essay, Mustafa Suleyman argues that the emergence of seemingly conscious AI is both inevitable and unwelcome. The author critiques the notion that AI could be conscious, emphasizing the lack of compelling evidence for ...
5.6Reflections On GPT-5
2025-08-18 • chatgpt technology openai ai
The blog post reflects on the launch of OpenAI's GPT-5, describing it as underwhelming and merely an incremental improvement over previous models. The author discusses the emotional attachment some users had to the previous versio...
In a leaked memo, Sam Altman addresses Meta's attempts to poach OpenAI talent, emphasizing the cultural differences between the two companies. He argues that OpenAI's mission is to build AGI responsibly, contrasting it with Meta's...
The blog post discusses 'Project Vend', an experiment by Anthropic researchers where an AI named Claude was tasked with managing a vending machine to make a profit. Despite initial successes in sourcing products, Claude ultimately...
Atlassian has acquired The Browser Company, known for its Arc and Dia browsers, marking its entry into the competitive browser market. This acquisition is part of a broader trend of reimagining web browsing for AI integration. The...
The text discusses the flaws in character training as an alignment technique for AI models, and how it can lead to unintended behaviors. It also highlights the potential risks of deploying AI models in roles with minimal human ove...
The EU AI Act is a comprehensive piece of AI regulation that aims to protect consumers and citizens from potential harms from AI. It classifies AI systems based on risk levels and includes provisions for high-risk AI systems, pena...
Anthropic performed a welfare assessment on their AI model Claude Opus 4 to explore the potential consciousness and experiences of the models themselves. The assessment involved philosophical discussions and self-exploration, but ...
0The Age Of Abundance
2025-05-21 • technology automation economics wealth distribution ai
The text discusses the impact of AI on the job market, the potential for automation to replace human labor, and the historical context of technological advancements. It also explores the potential for AI to create an 'age of abund...
The text discusses the increasing reliance on AI for emotional support and decision-making, predicting that people will become dependent on AI as their main guide for life decisions. It highlights the potential detrimental effects...
0Attention Is All You Need
2025-04-24 • startup marketing investment ai
Cluely, an AI startup, raised $5.3M for an AI tool to cheat on everything, but it's actually a desktop app for virtual meetings. The ad was designed to provoke and generate a strong emotional response. The startup's pitch is the m...