AI BreakfastAI Breakfast You read. We listen. Let us know what you think by replying to this email. 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. Watch the Recording Claude Fable 5 can now do 16% of freelance jobs at a professional level Anthropic has removed covert steganography code from Claude Code after developers discovered hidden Unicode markers used to catch Chinese competitors from distilling its data. The company is replacing this defensive infrastructure with behavioral fingerprinting and fake data injection. At the same time, Anthropic is tightening safety controls around its Fable 5 models. New safeguards now downgrade sensitive cybersecurity or biology-related requests to weaker models, while a new “Cyber Jailbreak Severity” framework scores AI exploits based on how dangerous and weaponizable they are. The business case for AI automation is accelerating. According to the Center for AI Safety’s latest Remote Labor Index, Claude Fable 5 now completes 16.1% of professional remote work projects, doubling the capability of its closest rival. However, this heavy reliance has exposed a massive vulnerability. Following a sudden government halt on Fable 5, a VentureBeat survey revealed that 67% of enterprises are already hedging their risk by shifting critical workflows to open-weight or self-hosted systems to protect themselves from future disruptions. On the development side, optimizing these deployments requires a shift in strategy. Anthropic cut the system prompt for Claude Code by 80% after discovering that its new Mythos-class models perform better with fewer instructions. Heavy prompting and rigid rules actually hurt the natural reasoning ability of these advanced models, prompting developers to move toward lighter, higher-level guidance instead. To manage autonomous agents, Anthropic launched Claude Enterprise spend controls, which track model usage and connect directly to Datadog. Meanwhile, Artifacts expanded to Pro and Max tiers, enabling live, self-updating code pages. Anthropic is also reportedly entering early planning with Samsung for proprietary silicon designed by former Tesla and OpenAI engineers to bypass expensive GPU limits. Finally, TestingCatalog published an early look at Claude Science in beta. The platform equips researchers with 60 specialized tools across genomics and cheminformatics to automate workflows, such as CRISPR screen design, on high-performance computing clusters. OpenAI might give the US government a 5% stake to fund citizen payouts OpenAI is trying to fix its relationship with both Washington and a worried public, and it might cost them a 5% stake in the company. CEO Sam Altman is floating a plan to hand that equity over to the US government before OpenAI goes public. The idea is to seed a public wealth fund, like Alaska’s oil-dividend model, to pay out citizens and ease anxiety over AI taking people's jobs. To top it off, Altman wants a US-led global AI watchdog modeled after the International Atomic Energy Agency to police advanced model training. But analysts warn this could backfire globally, leading other countries to demand their own piece of the pie and wrecking international data sovereignty. Meanwhile, a new research paper from OpenAI accidentally leaked an unannounced GPT-5.6 Pro lineup, giving us a clear look at how companies will actually deploy this tech. Instead of one giant, expensive model doing everything, OpenAI is building three specialized, task-specific variants: Luna Pro for speed, Terra Pro for massive data crunching, and Sol Pro for deep reasoning. The benchmarks show that by letting the mid-tier Terra Pro chew on a problem longer, it nearly matches the performance of the flagship Sol model at a fraction of the cost. Grok now lets developers build apps using just their voice xAI is turning Grok into a serious development tool. The company just rolled out Grok Connectors and a new Grok Build feature, pushing the chatbot into workflow automation. It can now turn visual designs into working code and generate architecture diagrams right inside FigJam. Plus, developers can now voice-code in real time using a new speech-to-text feature called Grok Voice, triggered by a simple /voice command. To force adoption, Tesla is clamping down on outside AI spending. Starting July 6, Tesla will cap employee AI budgets at $200 a week to manage soaring inference costs, but Grok usage is completely exempt. The move effectively pushes engineers away from preferred tools like Claude and into Musk's own ecosystem. Rumors swirled that SpaceX showed investors a prototype AI device powered by Qualcomm chips and xAI models that links directly to Starlink, though Musk denied the report. Meanwhile, X is also trying to lure broadcasters away from competitors by launching a revamped desktop livestream studio for Premium subscribers, backed by a fresh $1 million creator funding pool. Frontier models and product moves
Z.ai launches ZCode to take on Cursor and GitHub Copilot
Z.ai launches ZCode to take on Cursor and GitHub CopilotMeta claims its Watermelon model secretly caught up to GPT-5.5
Meta claims its Watermelon model secretly caught up to GPT-5.5Alexandr Wang promises a Muse Spark upgrade to rescue Meta's AI agents
Alexandr Wang promises a Muse Spark upgrade to rescue Meta's AI agentsGemini Omni Flash clinches the top video spot in Design Arena, outscoring Seedance Mini
Gemini Omni Flash clinches the top video spot in Design Arena, outscoring Seedance MiniArtificial Analysis: Claude Sonnet 5 nearly matches Fable 5 at 17x the cost
Artificial Analysis: Claude Sonnet 5 nearly matches Fable 5 at 17x the costBefore its shutdown, Fable 5 broke browser AI speed limits with Gemma
Before its shutdown, Fable 5 broke browser AI speed limits with GemmaAlibaba is giving away $5,000 to prove its new AI model beats Claude
Alibaba is giving away $5,000 to prove its new AI model beats ClaudeGoogle might be testing a Gemini Flash upgrade on LM Arena
Google might be testing a Gemini Flash upgrade on LM ArenaThe twilight of the chatbots — by Ethan Mollick
The twilight of the chatbots — by Ethan Mollick Agents and the agentic stackNew Alibaba AI framework drops token use from 884,000 to just 1,160
New Alibaba AI framework drops token use from 884,000 to just 1,160Vercel's Andrew Qu on why agents are a new kind of software
Vercel's Andrew Qu on why agents are a new kind of softwareHermes Agent's Judgment Release makes autonomous coding more trustworthy
Hermes Agent's Judgment Release makes autonomous coding more trustworthyNew benchmark from ByteDance reveals how AI agents learn over multi-day marathons
New benchmark from ByteDance reveals how AI agents learn over multi-day marathonsWhy enterprise executives still fall for the sleeper agent AI trap
Why enterprise executives still fall for the sleeper agent AI trapMicrosoft launches Frontier Company to save failing corporate AI experiments
Microsoft launches Frontier Company to save failing corporate AI experimentsCisco is rolling out AI agents to every single one of its 90,000 employees
Cisco is rolling out AI agents to every single one of its 90,000 employeesRobinhood debuts a layer-2 network to let AI agents trade global stocks
Robinhood debuts a layer-2 network to let AI agents trade global stocksLangChain releases OpenWiki to keep codebase docs fresh for AI agents
LangChain releases OpenWiki to keep codebase docs fresh for AI agentsGiving AI coding tools a senior design engineer's aesthetic judgment
Giving AI coding tools a senior design engineer's aesthetic judgmentThe quiet engineering war against probabilistic AI code generation
The quiet engineering war against probabilistic AI code generation Business, labor, and institutionsUS feds are actively hiring the person who will decide which AI models to ban in real-time
US feds are actively hiring the person who will decide which AI models to ban in real-timeEpoch AI tracks an unprecedented surge in critical software vulnerabilities
Epoch AI tracks an unprecedented surge in critical software vulnerabilitiesAI could slash the deficit by $2.2 trillion, but there is a catch
AI could slash the deficit by $2.2 trillion, but there is a catchInternal memo reveals Microsoft is merging Copilot apps to cut bloat (paywalled)
Internal memo reveals Microsoft is merging Copilot apps to cut bloat (paywalled)A new Cloudflare policy forces AI bots to pay web publishers
A new Cloudflare policy forces AI bots to pay web publishersPalantir CEO Alex Karp doesn't hold back as he rails against the AI industry
Palantir CEO Alex Karp doesn't hold back as he rails against the AI industryHere's how much more expensive your devices have gotten thanks to AI
Here's how much more expensive your devices have gotten thanks to AIApple weighs blacklisted Chinese memory to beat an AI component squeeze
Apple weighs blacklisted Chinese memory to beat an AI component squeezeNvidia offers to lease back its own unsold GPUs from small cloud startups
Nvidia offers to lease back its own unsold GPUs from small cloud startupsUS FTC says AI bias safeguards may run afoul of consumer law
US FTC says AI bias safeguards may run afoul of consumer lawSuffolk embeds AI engineers on active jobsites to curb construction friction
Suffolk embeds AI engineers on active jobsites to curb construction frictionAI grief videos turn mourning into a $390 service in South Korea (paywalled)
AI grief videos turn mourning into a $390 service in South Korea (paywalled)Meta experiments with prompt-to-play mini games in its new Pocket app
Meta experiments with prompt-to-play mini games in its new Pocket appNew NYU report builds the official playbook for tracking AI welfare
New NYU report builds the official playbook for tracking AI welfareWeave is actually shipping a home robot this year
Weave is actually shipping a home robot this yearLeaked video shows Microsoft's work on a lightweight Copilot OS built for AI
Leaked video shows Microsoft's work on a lightweight Copilot OS built for AIYou can now pre-order this $8,000 laundry-folding robot
You can now pre-order this $8,000 laundry-folding robot WF2026: Autoresearch and Keynotes ft. Anthropic, Google DeepMind, Amazon AGI, Sonar, Arena, Recursive Context.dev is the unified API to scrape, crawl, and extract structured, LLM-ready web data in minutes. Acti is an agentic mobile keyboard that executes commands, searches, and workflows without leaving your conversation. Humalike delivers the behavioral APIs and social intelligence infrastructure needed to build genuinely humanlike, proactive AI agents. Macro unifies your entire workflow by combining email, docs, CRM, and code into one workspace with shared team-level memory. Docubix is a RAG platform that indexes unstructured documents into source-backed, streaming AI assistants via a single API. Thank you for reading today’s edition. Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter. Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on X! Thinking of starting your own newsletter? AI Breakfast readers who sign up with Beehiiv receive a 14-day free trial and 20% off for 3 months. Keep Reading AI Breakfast Curated weekly analysis of the latest AI projects, products, and news