4. Artificial Intelligence

AU/NZ: Anthropic makes Fable 5 a permanent offering 

👓 1 minute read

After weeks of uncertainty, Anthropic has confirmed that its Fable 5 model will remain permanently available to Max and Team Premium subscribers, although usage will be limited to half of each plan's allowance. Lower-tier subscribers will receive a one-off US$100 credit before moving to a pay-per-use model.  

The decision follows three postponed deadlines, strong customer criticism over changing access arrangements, and increasing competition from OpenAI's GPT-5.6 Sol and Moonshot's Kimi K3.  

Anthropic said demand had been difficult to predict and that it is investing in additional computing capacity, but the episode highlights the growing challenge AI developers face in balancing frontier model performance with the infrastructure needed to support widespread access 

AU/NZ: Mastering Claude Fable 5 for high-value work 

👓 1 minute read

Samantha McLean says that Anthropic's Claude Fable 5 should be treated as a strategic adviser rather than an everyday writing assistant. With the model moving to pay-per-use pricing, she suggests agents should reserve it for complex, high-value decisions such as business planning, market analysis, operating reviews and long-term client nurture strategies, while leaving routine tasks like listing copy, emails and social media posts to lower-cost models. 

Her most useful tips are: 

Turn your week into an operating plan. McLean suggests pasting listing notes, OFI feedback, emails, WhatsApp conversations and to-do lists into Fable and asking it to produce a structured operating plan. The emphasis is on identifying commitments, decisions, conflicts and the five highest-priority actions for the following week.  

Stress-test major business decisions. Rather than asking AI what to do, she recommends having it interview you about a decision (such as moving to an auction-first strategy), challenge vague thinking and present a recommendation together with the strongest argument against it.  

Research new markets more rigorously. She provides a framework for evaluating expansion into a suburb or region, requiring the AI to distinguish evidence from inference, cite sources and even critique its own conclusions before presenting a recommendation.  

Develop market commentary from real observations. Rather than asking AI to "write a market update", she suggests feeding it raw notes from buyer conversations, OFIs and vendor meetings so it identifies the underlying story before any drafting begins.  

Separate strategy from production. One of her better recommendations is to use Fable to decide what should be produced, then hand those briefs to a cheaper model (or a person) to create the actual listing copy, emails and social posts.  

Read the full article to access prompts.

AU/NZ: NotebookLM becomes Gemini Notebook 

👓 1 minute read

Google has rebranded NotebookLM as Gemini Notebook, with the research assistant now becoming more closely integrated with the wider Google ecosystem. Existing notebooks will continue to work as before, but users will soon be able to access them directly within the Gemini app and, later, through AI Mode in Google Search.  

The more significant change is under the hood.  

Google is introducing a secure cloud computer that allows Gemini Notebook to write and execute code while analysing information contained within your source documents. For estate agents, this opens the door to more sophisticated analysis of market data, sales results and research material without leaving the notebook environment. The capability is rolling out first to Google AI Ultra and eligible Workspace business customers, before expanding to Pro users in the coming weeks.