Ted Chiang, a renowned science fiction writer, has reaffirmed his stance that artificial intelligence is not conscious. This statement is consistent with his previous interviews and essays, where he emphasized the limitations of current AI systems. Chiang's argument is not just about the technology itself, but also about the language we use to describe it, often perpetuating misconceptions about its capabilities.
As we reported on June 3, discussions around AI and its potential impact on society have been gaining momentum, with voices like Bernie Sanders warning about the dangers of unchecked AI development. Chiang's perspective is significant, as it highlights the need for a more nuanced understanding of AI and its potential applications. By recognizing the limitations of current AI systems, we can temper our fears and expectations, and focus on developing more realistic and beneficial technologies.
Looking ahead, it will be interesting to see how Chiang's views influence the ongoing debate about AI and consciousness. As AI continues to evolve, it is crucial to have informed discussions about its potential implications, and Chiang's insights can help shape a more informed and balanced conversation. With his unique blend of literary and technical expertise, Chiang is well-positioned to contribute to the development of a more thoughtful and responsible AI ecosystem.
A recent post from an AI practitioner reveals their approach to working with AI in 2026, using a laptop's integrated GPU and renting four Frontier GPUs. This setup allows them to engage with AI while maintaining control and accountability, with a human-in-the-loop approach and their name on every output.
This matters because it highlights the growing need for professionals to adapt to AI, despite concerns around copyright, energy consumption, and job displacement. As we reported on June 2, OpenAI's new Codex tools are expanding AI's reach into white-collar work, making it essential for individuals to find ways to work with AI effectively.
As the AI landscape continues to evolve, it will be interesting to watch how professionals balance the benefits of AI with the potential risks and challenges. With the development of more efficient hardware, such as NPUs and improved iGPUs, and the growth of AI rental services, we can expect to see more innovative approaches to AI adoption in the coming months.
Mnemo, a local-first AI memory layer, has been introduced for use with any Large Language Model (LLM). This innovation allows for persistent knowledge graphs, entity extraction, and semantic retrieval without relying on cloud services. Most LLMs currently forget conversations once they end, but Mnemo acts as a sidecar service, watching and extracting information from every conversation.
This development matters because it addresses a significant limitation in current LLM technology. By enabling LLMs to retain memory of past conversations, Mnemo has the potential to significantly enhance their ability to learn and interact with users. This could lead to more personalized and effective AI-powered applications across various industries.
As Mnemo continues to evolve, it will be important to watch how it integrates with different LLM backends, such as OpenAI, Anthropic, and Ollama. Additionally, the project's use of Rust, SQLite, and petgraph suggests a focus on efficiency and scalability, which will be crucial as it is adopted by a wider range of users. With its open-source availability on GitHub, the community can contribute to and shape the future of Mnemo, potentially leading to new breakthroughs in AI memory and cognition.
As we reported on the development of embedding-based routing, a new update has been released, marking the shipment of Phase 2. The latest post, "Phase 2 Shipped: 5 Things I Got Wrong About Embedding-Based Routing," serves as a follow-up to "Teaching an AI to Pick Its Own Brain," where the author outlined a plan to improve the technology.
This update matters because choosing the right embedding model is crucial for efficient feature shipping and retrieval optimization. As highlighted in a Medium post, selecting the wrong model can lead to significant time and resource waste, with teams potentially spending months on optimization instead of development. The rise of Industry 4.0 is also driving the deployment of embedded boards and modules in manufacturing facilities, making the development of efficient embedding-based routing systems even more critical.
Looking ahead, it will be interesting to see how the latest advancements in embedding models, such as the nomic-embed-text-v2-moe, impact the field. This model has shown high performance in multilingual retrieval, outperforming larger models. As the technology continues to evolve, we can expect to see improved efficiency and capabilities in AI-powered systems, particularly in industries adopting Industry 4.0 technologies.
Apple has started requiring age verification for App Store use in Texas, following a state law passed earlier this year. This law mandates app stores to verify users' ages, and Apple is now complying with the regulation. The move is significant as it sets a precedent for age verification in the tech industry, potentially influencing other states or countries to introduce similar laws.
This development matters because it raises questions about data privacy and the potential impact on app developers. As Apple and other app stores implement age verification, they must balance user protection with the need to safeguard personal data. The Texas law may also have implications for the use of Large Language Models (LLMs) and AI agents in app development, as discussed in our previous reports on WarAgent and Hyper.
As the situation unfolds, we will be watching how Apple's age verification process is implemented and received by users. We will also be monitoring whether other states or countries follow Texas's lead in introducing similar laws, and how this affects the broader tech industry, including the development of AI-powered apps and services.
The rising cost of human developers has become a significant concern for companies, making them increasingly reliant on chatbots and AI tools. As we reported on June 3, the integration of Artificial Intelligence in various industries, including human resources and development, is transforming the way organizations work. However, this shift also highlights the expense of hiring human talent, particularly in fields like tech and development.
The preference for chatbots over human developers is not only driven by cost but also by the efficiency and availability of these AI tools. With the ability to simulate human-like conversations, chatbots have become a popular choice for addressing customer queries and streamlining processes. As companies continue to adopt AI solutions, the demand for human developers may decrease, making it challenging for professionals in this field to find employment.
As the AI industry continues to evolve, it is essential to address the concerns surrounding the replacement of human workers with chatbots and AI tools. The focus should be on developing AI solutions that complement human capabilities, rather than replacing them entirely. With the emergence of new AI tools and technologies, companies must reassess their hiring strategies and consider the long-term implications of relying solely on chatbots and AI-powered solutions.
OpenAI and Anthropic have signed a letter, along with other top AI CEOs, urging Congress to pass legislation that prevents the development of biological weapons using AI. This move comes as a proactive measure to address concerns about the potential misuse of AI systems. As we reported on June 3, Anthropic's valuation has skyrocketed to $380 billion, and OpenAI is facing a lawsuit in Florida over violent incidents, highlighting the need for regulation in the AI industry.
The letter, signed by Google DeepMind's Demis Hassabis, OpenAI's Sam Altman, and Anthropic's Dario Amodei, among others, demonstrates a unified effort by AI leaders to acknowledge and mitigate the risks associated with AI development. This is not the first time these companies have taken steps to address safety concerns - in March, Anthropic and OpenAI hired weapons experts to prevent their AI systems from being used to create weapons of mass destruction.
As the AI industry continues to evolve, it is crucial to watch how governments and regulatory bodies respond to these concerns. The recruitment of experts and the signing of this letter indicate a willingness from AI companies to work towards safety and regulation, but concrete legislation is needed to prevent the misuse of AI. The next steps will be crucial in determining the future of AI development and its potential impact on global security.
The hidden cost of AI agents has become a pressing concern for developers, as small decisions can lead to significant expenses over time. As we reported on June 3, OpenAI has been offering UK banks cyber tool access, while Anthropic limits Mythos, highlighting the growing need for efficient AI tool management. A recent report reveals that costs rise from longer tasks, retries, and hidden agent calls, resulting in surprise bills that can amount to $100k/yr.
The issue lies in the lack of observability and monitoring of AI agent deployments, making it challenging to track token usage and costs. To avoid such expenses, developers can set limits, use smaller models when possible, cache results, and monitor usage closely. Implementing observability through OpenTelemetry integration, tool call tracing, and session replay can also help identify areas of inefficiency. Furthermore, cost tracking and breakdowns can be achieved by separating token usage and costs into input, output, and other categories.
As the development of reliable AI agents continues, it is essential to focus on concurrency, retries, and timeouts. The use of a single-owner pattern with AbortSignal, deadline budgets, and jittered retries can help fix issues such as Promise.race leaks, which can lead to billing leaks in AI agents. Developers should watch for updates on cost monitoring and observability tools, as well as best practices for building reliable AI agents in TypeScript, to stay ahead of the curve and avoid hidden costs.
Anthropic's latest flagship model, Claude Opus 4.8, is facing intense criticism just as the company prepares for its initial public offering (IPO). The model, touted for its enhanced performance in coding and professional tasks, has been plagued by issues of identity confusion and astronomical costs. This backlash comes at a crucial time, as Anthropic has recently filed its IPO application with the US Securities and Exchange Commission (SEC), seeking to raise $65 billion at a valuation of $965 billion, surpassing OpenAI.
The controversy surrounding Claude Opus 4.8 matters because it raises questions about the reliability and affordability of Anthropic's AI technology, potentially impacting investor confidence in the company's upcoming IPO. As Anthropic aims to become a leader in the AI market, the performance and reputation of its flagship model are crucial to its success.
As the situation unfolds, it will be essential to watch how Anthropic responds to the criticism and whether the company can address the issues with Claude Opus 4.8. The outcome may have significant implications for Anthropic's IPO and its position in the competitive AI landscape. With the company's valuation and funding at stake, Anthropic must swiftly resolve the problems with its flagship model to maintain investor trust and achieve its ambitious goals.
A recent experiment has shed light on the vulnerabilities of Large Language Models (LLMs) in app security. Kasra, a developer, built a deliberately vulnerable app and spent $1,500 to test if LLMs could hack it. The results, documented on Kasra's blog, highlight the potential risks of relying on LLMs for critical decisions.
This experiment matters because it demonstrates the real-world implications of LLM vulnerabilities, which have been discussed in the context of AI security. As we previously reported, Apple has been involved in an AI lawsuit, and the topic of LLM security has been gaining attention. The fact that Kasra's app was successfully hacked by LLMs raises concerns about the security of AI-powered applications, particularly those handling sensitive information.
As the use of LLMs becomes more widespread, it is essential to watch for developments in AI security and the measures being taken to prevent attacks. The GitHub repository "Vulnerable LLMs" and guides like "How to Build Secure LLM Apps and Prevent Attacks" on LinkedIn provide resources for developers to learn about LLM security risks and prevention strategies. The experiment's findings will likely contribute to the ongoing conversation about the importance of securing LLMs and the need for robust testing and validation.
OpenAI CEO Sam Altman is set to testify before US lawmakers, advocating against proposals that require AI developers to obtain government approval before releasing new models. This comes as part of a broader effort to shape regulation of the technology. As we reported on June 3, Florida's lawsuit against OpenAI and Altman has brought attention to the need for regulation, with many calling for stricter oversight of AI development.
The move matters because it highlights the ongoing debate over how to regulate AI. While some argue that government approval is necessary to prevent harm, others claim it could stifle innovation. Altman's testimony is significant, as it reflects the industry's concerns about over-regulation. With AI models like ChatGPT and Gemma 4 12B becoming increasingly powerful, the need for effective regulation is growing.
As the Senate hearing unfolds, it will be crucial to watch how lawmakers respond to Altman's testimony. Will they heed his warnings about over-regulation, or will they push for stricter controls? The outcome will have significant implications for the future of AI development in the US, and the world will be watching to see how this plays out.
Warner Bros has partnered with OpenAI to integrate movie ticket purchasing into ChatGPT, transforming the popular chatbot into a movie ticket machine. This move marks a significant expansion of ChatGPT's capabilities, leveraging its vast user base to drive sales and revenue. As we reported on June 4, OpenAI's CEO Sam Altman is urging US lawmakers not to require AI model approvals, indicating the company's push for innovation and growth.
This development matters because it showcases the potential of AI-powered platforms to disrupt traditional industries, such as entertainment and commerce. By integrating e-commerce functionality, OpenAI is diversifying its revenue streams and positioning ChatGPT as a computational marketplace. This strategic pivot is likely a response to the company's financial challenges, including losses on its ChatGPT Pro plan.
As this partnership unfolds, it will be interesting to watch how users respond to the new feature and whether other companies follow suit. Will this move pave the way for further integration of AI-powered commerce in various industries, or will it raise concerns about data privacy and security? The success of this initiative will likely depend on OpenAI's ability to balance innovation with user trust and regulatory compliance.
A developer has created a circuit breaker for LLM agents after witnessing someone lose $200 overnight due to looping runs on their first question. This incident highlights the financial risks associated with AI agents, a concern we've been tracking since reporting on a vulnerable app being hacked by LLMs. The newly built circuit breaker, called AgentCircuit, is an open-source decorator that provides loop detection, auto-repair, output validation, and budget control.
This development matters because it addresses a critical need for safety and reliability in AI agents. As we've seen in previous cases, such as the $437 API bill incident, the lack of safeguards can lead to significant financial losses. AgentCircuit's features, including fuse, sentinel, medic, and budget controls, can help prevent such incidents and provide a more robust framework for AI agent development.
As the use of LLMs and AI agents continues to grow, the importance of circuit breakers and safety mechanisms will only increase. We can expect to see more developments in this area, with a focus on hardwiring safety into the internal workings of LLMs and AI agents. The creation of AgentCircuit is a significant step forward, and its open-source nature will likely lead to further innovation and refinement in the field.
Adobe's creative tools are now integrated with Gemini, a significant development in the AI landscape. This move enables seamless collaboration between Adobe's suite of creative applications and Google's Gemini AI model. As we reported on the growing presence of AI in creative fields, this integration is a natural next step, allowing designers and artists to leverage the power of AI in their work.
The integration matters because it bridges the gap between human creativity and AI-driven innovation. By combining Adobe's industry-standard creative tools with Gemini's advanced AI capabilities, users can tap into new levels of productivity and inspiration. This partnership has the potential to revolutionize the way we approach design, art, and content creation.
As the creative industry continues to evolve, it's essential to watch how this integration unfolds. We can expect to see new features and functionalities emerge, further blurring the lines between human and machine creativity. With Adobe and Google at the forefront of this development, it will be interesting to see how other industry players respond and adapt to this new landscape.
Anthropic, the developer of AI model Claude, has confidentially submitted a draft S-1 registration statement to the US Securities and Exchange Commission, marking the beginning of its initial public offering (IPO) process. This move could potentially allow Anthropic to go public after the SEC review is complete, posing a challenge to OpenAI, a leading player in the AI market.
The IPO preparation is significant as it underscores Anthropic's ambition to expand its presence in the AI landscape, particularly with its agentic AI model Claude, which is seen as a competitor to OpenAI's ChatGPT. As we reported on June 4, Anthropic has been facing criticism over its flagship model Claude Opus 4.8, but this development suggests the company is pushing forward with its growth plans.
What to watch next is how the SEC review process unfolds and how Anthropic's IPO plans will impact the AI market, particularly its relationship with OpenAI. With OpenAI's CEO Sam Altman recently discussing the AI revolution and the role of AI agents, the competition between these two companies is likely to heat up, leading to further innovations and developments in the AI space.
OpenAI has announced the release of six role-based plugins and Sites for its Codex platform, marking a significant step towards making AI agents more accessible to non-developers. This move is part of the company's efforts to democratize access to AI technology, allowing a broader range of users to leverage the power of AI in their daily work.
As we reported on June 4, OpenAI has been actively exploring various applications of its technology, including integrating ChatGPT into a movie ticket machine and urging US lawmakers not to require AI model approvals. The latest development with Codex plugins and Sites is a natural extension of this push, enabling users to tap into the potential of AI without requiring extensive coding knowledge.
What matters here is the potential for AI to become an integral part of various workflows and roles, beyond just development. With Codex now available on AWS and offering a range of plugins, users can expect to see more seamless integration of AI into their daily tasks. As OpenAI's CEO Sam Altman has emphasized, the AI revolution is here to stay, and this latest move is a testament to the company's commitment to making AI more accessible and user-friendly. As the landscape continues to evolve, it will be interesting to watch how non-developers adopt and utilize these new AI-powered tools, and what impact this has on the broader AI ecosystem.
The University of Chicago is taking a significant step in its artificial intelligence initiative by partnering with Anthropic to provide Claude Enterprise to its entire campus community. As of July, all academics and staff will have access to Claude Chat, Cowork, and Code, with students gaining access before the fall term. This move is part of the university's broader approach to AI, which aims to harness the technology's capabilities while addressing concerns around information security, privacy, compliance, and academic integrity.
This development matters because it underscores the growing importance of AI in higher education. By providing its community with access to cutting-edge AI tools, the University of Chicago is positioning itself at the forefront of this trend. The partnership also reflects the institution's commitment to responsible AI adoption, recognizing the need for careful consideration of the technology's potential risks and challenges.
As the University of Chicago rolls out Claude Enterprise, it will be worth watching how the campus community utilizes these tools and how they impact academic and administrative work. The university's AI initiative, launched in 2024, has already funded 15 proposals spanning research and education, and this latest development is likely to further accelerate innovation and collaboration across disciplines. With President Paul Alivisatos at the helm, the university is poised to become a leader in AI research and education, and its experiences will likely inform AI adoption strategies at other institutions.
Benchmark results are in for running Qwen3.6-35B-A3B on two 8-year-old GTX 1080 Ti cards, achieving approximately 20 tokens per second. This is only an 18% increase over CPU-only performance on an i9-14900K, which reaches around 17 tokens per second. The sparse MoE design of Qwen3.6-35B-A3B enables it to run on older hardware, but a second GPU does not double the speed as one might expect.
This matters because it shows that even older GPUs can be used to run large AI models, albeit with some limitations. The Qwen3.6-35B-A3B model's 3B-active design is what makes it possible to run on two GTX 1080 Ti cards. This has implications for those looking to run AI models on local hardware without breaking the bank on the latest GPUs.
As we look to the future, it will be interesting to see how other large AI models perform on older hardware and whether similar sparse MoE designs can be used to make them more accessible. Additionally, the community's experimentation with different models and hardware configurations, as seen in the Qwen 3.5 and 3.6 comparisons, will continue to push the boundaries of what is possible with local AI setups.
As the AI landscape continues to evolve, a fundamental question has resurfaced: what exactly is artificial intelligence? This inquiry comes on the heels of recent developments, including Warner Bros and OpenAI's integration of ChatGPT into a movie ticket machine, as reported on June 4. The term 'artificial intelligence' has been widely used, but its meaning and implications are not universally understood.
The ambiguity surrounding AI stems from its rapid growth and diverse applications, ranging from machine learning to natural language processing. Experts and enthusiasts alike are reexamining the concept, with some arguing that the term 'artificial intelligence' is a misnomer. As Chris News notes, AI exploration features software like the Google Brain project, while Visyu Trip defines AI as intelligence displayed by models, distinct from human intelligence.
As the conversation around AI continues to unfold, it is essential to watch for further clarification on the term's meaning and its implications for the industry. With companies like Anthropic and OpenAI leading the charge in AI development, a deeper understanding of the technology is crucial for investors, consumers, and regulators. As the AI money race intensifies, a clear definition of artificial intelligence will be vital in shaping the future of this rapidly evolving field.
Apple Music Classical has announced a new partnership with London's prestigious Wigmore Hall, a renowned center for classical music. This collaboration is set to bring exclusive content to Apple Music Classical subscribers, further enhancing the platform's offerings. As we reported on June 3, Apple has been actively expanding its music services, with Nintendo Music becoming available on iPad and CarPlay, and Apple TV impressing with its streaming capabilities.
This partnership matters because it underscores Apple's commitment to providing high-quality classical music content to its users. Wigmore Hall's rich history and reputation for showcasing exceptional talent will undoubtedly elevate Apple Music Classical's portfolio. The move also highlights the growing importance of strategic partnerships in the music streaming industry, as seen in Houston Grand Opera's recent collaboration with the London Symphony Orchestra.
As this partnership unfolds, it will be interesting to watch how Apple Music Classical leverages Wigmore Hall's expertise to curate unique experiences for its subscribers. With Wigmore Hall's 125th Anniversary celebrations underway, featuring premieres and exclusive performances, Apple Music Classical users can expect access to a treasure trove of classical music content. The upcoming summer season of classical music festivals in the UK will also be worth watching, as Apple Music Classical may potentially offer exclusive coverage or promotions tied to these events.
