IBM cosies up with Anthropic while Microsoft pushes into AI healthcare: the news that caught our attention this week
IBM strikes AI partnership with Anthropic
IBM has teamed up with Anthropic to integrate Anthropic's Claude family of AI models into its enterprise software development portfolio. The two companies have also created a guide on how enterprises can build, deploy, and maintain enterprise-grade AI agents.
IBM says Claude has already been added to its AI-first integrated development environment (IDE). The IDE is available in preview to select clients, and IBM reported that early internal testing among 6,000 users has shown average productivity gains of about 45% in coding tasks, while maintaining code quality and governance.
Dinesh Nirmal, SVP, Software at IBM, said: “This partnership enhances our software portfolio with advanced AI capabilities while maintaining the governance, security, and reliability that our clients have come to expect. We’re giving development teams AI that fits how enterprises work, not experimental tools that create new risks.”
The Motley Fool commented that the deal makes sense for both companies: Anthropic can leverage IBM's deep corporate relationships and get its product in front of more users, while IBM stays up to date with AI advances, and can give its customers access to the latest AI models.
The deal has some implications that product managers might want to consider. Embedding Claude into an IDE shows AI evolving from simply assisting developers to performing complex, multi-step tasks autonomously. It’s a deal that will also trouble OpenAI, whose positioning in enterprise software is less entrenched than IBM’s client relationships. The deal also offers enterprises an alternative to the troubled OpenAI/ Microsoft partnership. The jointly-published guide on agent design also points to Anthropic wanting to set standards for how enterprises should architect AI agents in regulated settings.
Financial terms of the agreement have not been disclosed, and it’s not yet clear how the partnership will affect the development of IBM’s own AI models.
Google expands vibe coding tool availability
Google has expanded the release of its vibe-coding tool Opal and it is now available as a public beta in a further 15 countries including Canada, India, and Japan, The tool was initially launched in the US in July.
The vibe coding market is seeing explosive growth, with companies like Lovable and Replit receiving significant funding rounds. A recent Gartner study, Why Vibe Coding Needs To Be Taken Seriously, estimates that 40% of new enterprise production software will be created with vibe-coding techniques and tools by 2028.
How TikTok keeps its users addicted
The Washington Post has used a novel data donation strategy to investigate why TikTok, with a famously opaque algorithm, is so addictive, keeping its subscribers in-app and scrolling. The publication has found that the social platform uses an AI-driven algorithm to create a highly personalised "For You" page, which constantly tailors content to individual user preferences and maximises engagement through “dopamine-releasing, addictive feedback loops”.
Over 800 US TikTok users shared their data with the publication which found that after a month, over three-quarters of the users who had typically spent 30 minutes per day on TikTok were now spending, on average, nearly twice that time on the app each day.
In summary, The Washington Post identified several tactics used by the algorithm to ensure a TikTok habit forms quickly: for example, it provides a constant stream of new content by serving up videos from accounts users don’t follow, and doesn’t rely on users to state their interests, instead learning by observing their behaviour.
According to TikTok internal documents, made public as part of a multi-state lawsuit against the company, it takes just 260 videos, which might only take 35 minutes, to form a habit on the app.
European Commission launches initiatives to speed up AI uptake
The European Commission has launched two strategies to speed up AI uptake in European industry and science.
The Apply AI Strategy aims to drive adoption of AI across strategic and public sectors, while the AI in Science Strategy positions the EU as a hub for AI-driven scientific innovation. At its centre is RAISE – the Resource for AI Science in Europe, a virtual European institute to pool and coordinate AI resources for developing AI and applying it in science.
With a budget of about €1 billion, the Apply AI Strategy measures include establishing AI-powered advanced screening centres for healthcare and supporting the development of frontier models and agentic AI tailored to sectors such as manufacturing, environment and pharmaceuticals. The AI in Science Strategy includes measures to attract talent, increased investment in compute power, increased funding for research, and support for scientists to identify strategic data gaps and gather, curate and integrate the datasets needed for AI in science.
Microsoft partners with Harvard Medical School in AI healthcare push
Microsoft has partnered with Harvard Medical School as part of its push into AI-powered healthcare. Microsoft is paying a licensing fee for access to the medical school’s consumer-focused medical content in order to improve the reliability of Copilot’s healthcare-related answers and related services by grounding them in clinically reviewed, authoritative sources.
The partnership is seen as part of Microsoft’s drive to reduce its reliance on OpenAI and a move to build trust in Microsoft's AI offerings. Newsletter Implicator.AI reported that Dominic King, vice president of health at Microsoft AI, said that Copilot should align more closely with what users might hear from medical practitioners. The newsletter also commented that the deal is the clearest signal yet that Microsoft’s AI division is “building vertical-specific moats rather than chasing general-purpose scale”.
