Cognitive Warmup. If you hear about Anthropic more than once in our conversation this week, it’s solely because a few things need to be clarified about the AI company. First, a federal appeals court in Washington rejected a petition by the U.S. War Department to block the “supply-chain risk” designation. This means Anthropic’s defense contracts will remain restricted. (The second, in our conversation later this week.)
This comes after the company in February refused to allow government use of cloud AI for fully autonomous weapons and home surveillance applications. Anthropic says current frontier AI systems are not reliable enough to make life-or-death decisions without human oversight.
Dario Amodei said at the time, “It is the department’s prerogative to select the contractors that best match its vision. But given the substantial value that Anthropic’s technology provides to our armed forces, we hope they will reconsider.”
Gemini’s important notebook
I must say: Google’s legacy approach to organization is much appreciated. Gemini is now getting an update that expands the scope of integration into Google’s other important AI app NotebookLM, adding Notebook Core to the Gemini experience – from managing notes for exams to tracking workflows for complex projects or even a new hobby. These are being rolled out first to subscribers and then to free-tier users in the coming weeks.
Rebecca Zapfel, senior product manager at Google DeepMind, says, “Starting with Gemini, think of notebooks as personal knowledge bases shared across Google products. They give you a dedicated space to organize your chats and files, and because they sync with NotebookLM, you can unlock even more efficient workflows directly from Gemini.”
The basic objective of the Gemini-NotebookLM integration will be to keep all data, research and conversations neatly organized with seamless sync across both the AI apps.
Learn a New Word: Cognitive Dedication
If your brain hasn’t become completely useless after handing over all the thinking to an AI chatbot, you might want to take the paper by Wharton School researchers Steven Shaw and Gideon Nave a little more seriously.
Shaw and Neave provided 1,372 people with access to a chatbot for help, a test adapter from the Cognitive Reflection Test. Crucial to this is the three-system theory, which extends Daniel Kahneman’s dual-process model of cognition (“thinking, fast and slow”) to add a third system, which is artificial intelligence.
“We show that people not only use System 3 to aid reasoning, but also often surrender to its output – whether correct or flawed. This cognitive surrender reflects the value and integration of System 3, but also highlights the vulnerability of System 3 use. Just as System 1-driven inference gives rise to systematic bias, System 3 has differential cognitive shortcomings that challenge decision makers and society at large. Will give,” the researchers concluded.
Now humans will have to answer some questions. What happens to decisions that are not made by our brains? What happens to effort and intuition when thinking is done by a machine or algorithm? And last but not least, how can we preserve elements of agency or reflection if we hand over our thinking to chatbots? I would like to add one more thing here – how lazy will we become?
legend of the cloud mythos
In my analysis a few days ago, I noticed an irony that lies behind the entire situation. As for an apparently super-powerful AI model that leaked through Anthropic’s content management system, there are claims that it will fix as-yet-unknown and soon-to-be-known vulnerabilities in all the world’s software.
This may or may not be the case (and it’s very likely here), but AI companies continue to weave exceptionally effective narratives, hoping no one will notice.
Anyway, Anthropic claims the Cloud Mythos is so powerful, they won’t make this model generally available. Instead, it provides the basis for something called Project Glasswing to secure every other software. Partners include Apple Inc., Google, Microsoft Corp., Amazon Web Services, Nvidia Corp. and Cisco Inc. Are included.
Anthropic says they will contribute up to $100 million in usage credits for Mythos Preview during these efforts. After that, Cloud Mythos will be available to preview participants at $25/$125 per million input/output tokens. It’s important to note the business structure being built here: The Claude mythos is still a product in Anthropic’s scheme of things – scarcity, capabilities, and so premium pricing strategies will be at work.
Neural Dispatch is your weekly guide to the rapidly evolving landscape of artificial intelligence. Would you like this newsletter sent to your inbox? Subscribe here.
