Sponsored by
News of the day
1. Meta launches Muse Spark 1.2, a coding-focused AI model with significant improvements in code generation, debugging, and developer workflows. → Read more
2. Former Spotify engineers launch Malachyte, raising $10M to apply their AI recommendation tech to e-commerce, focusing on real-time intent. → Read more
3. AI lab Mirendil partners with Google Cloud for $100M+ deal, securing compute for self-improving AI research and development. → Read more
4. Intel's SuperClaw beta explores hybrid AI, enabling AI agents to process sensitive data and tasks locally on PCs instead of solely relying on the cloud. → Read more
Our take
Hi Dotikers!
Meta is finally entering the coding agent wars. Muse Code, released yesterday in beta, is a terminal agent that plans, writes and validates code across entire repos, powered by Muse Spark 1.2, a model co-trained with its own harness. On paper, this is catch-up: Claude Code and Codex have owned this space for over a year, and the original Muse Spark was lagging precisely on agentic coding benchmarks. Simon Willison sees it above all as confirmation of a deeper trend: the capability that truly matters today is long-sequence agentic tool calling. Meta had to build its own agent just to train its model properly.
But the real story is the pricing. Two prices for the same model: 1.25 dollars per million input tokens for the standard version, 0.10 dollars if you agree to let Meta use your data to improve its models. A discount of over 90 percent. Real, verifiable coding trajectories are the most valuable data around right now, and Meta kindly offers to let you pay your bill with yours. After twenty years of monetizing your vacation photos, the house knows the recipe, it just changed the dish.
This launch echoes the bombshell from the day before: Jeff Dean is leaving Google after 27 years to found Discovery Loop, a startup that wants to automate the scientific experimental loop itself, hypothesis, experiment, evaluation, thousands of times in parallel. Meta is automating the developer's loop, Dean is automating the researcher's. The message is the same: value no longer lives in the model alone, but in the full loop you hand over to it. And those who own the traces of these loops hold the next competitive advantage.
Alex.
Usage-Based Pricing Is Here. Is Your Finance Team Ready?
More B2B companies are moving to usage-based and hybrid pricing — and finance teams are feeling it. Revenue recognition gets messier, forecasting gets harder, and the manual work compounds fast.
Tabs and PwC teamed up to break it down. In this on-demand session, Rebecca Schwartz and Amit Dhir share how leading finance teams are handling the operational reality of dynamic pricing models — and where AI fits in.
Watch the recording for concrete examples, a practical rev rec framework, and a clear-eyed look at what it takes to scale without the overhead.
If your team is navigating this shift, this session is worth your time.
Meme of the day
