Future Bytes

Trust is becoming a competitive advantage in AI. Claude surged as users switched from ChatGPT, even as OpenAI launched a powerful new model. Magnus Oxenwaldt breaks down why values, portability, and governance are quickly becoming key factors in choosing an AI provider. 

• Claude overtakes ChatGPT in the App Store. 
• OpenAI launches GPT-5.4 with computer-use capabilities. 
• Deloitte warns most companies lack AI agent governance. 
• A lawsuit involving Gemini raises new safety questions. 

Hosted on Acast. See acast.com/privacy for more information.

Hosted on Acast. See acast.com/privacy for more information.

In this episode of Future Bytes, host Magnus Oxenwaldt speaks with Luca Sanitelli, Senior Partner Solutions Engineer, from Shopify about AI-driven discovery, infrastructure and why clean, machine-readable product data is becoming the new competitive edge. 

Links:

Shopify.com

Find Luca´s LinkedIn here

Hosted on Acast. See acast.com/privacy for more information.

Accenture now ties promotions to AI tool usage. Anthropic cuts model costs to one-fifth. Every enterprise surveyed plan to expand AI agents. Regulation is moving fast across US states. The risk in 2026 is no longer moving too fast. It’s standing still. 

Highlights for week 9: 

• Accenture ties senior promotions to measurable AI tool usage  
• Anthropic’s Claude Sonnet 4.6 delivers near-flagship capability at one-fifth the former cost. 
• 100% of surveyed enterprises plan to expand agentic AI this year  
• US states advance AI regulation as global adoption accelerates  

Hosted on Acast. See acast.com/privacy for more information.

This time, host Magnus Oxenwaldt examines a recent UC Berkeley study that links heavy AI use to rising burnout, as saved time quickly turns into more work. 

A familiar pattern is emerging across industries. Many firms are reducing headcount based on expected AI gains. OpenAI has introduced ads in ChatGPT's free tier. A global safety report calls for layered controls as vendor safety teams shift. The tools work. The real question is how we use them. 

Highlights for week 8: 

• Heavy AI use increases speed but also raises the risk of burnout. 
• Workforce cuts are being made based on expected, not yet proven, AI gains. 
• Ad-supported AI models introduce new incentive questions for enterprises. 
• Safety experts call for layered controls as vendor teams evolve. 

Hosted on Acast. See acast.com/privacy for more information.

A simple AI plugin wiped billions off legal tech valuations. Agent platforms moved from demos to production. Apple set a new standard for how software is built. The signal is clear: AI is no longer assisting work. It is starting to do it. 

Host Magnus Oxenwaldt explains the meaning behind the headlines.

Top stories for week 7:

• A Claude contract-review plugin triggered a 10-20% sell-off across major legal software firms. 
• Enterprise AI agents move into production: OpenAI’s Frontier platform lets agents operate across multiple enterprise systems, not inside one app. 
• Apple redefines how software is built. Xcode now supports autonomous AI agents that write, test, and verify code end to end. 
• Models get cheaper and more reliable. New releases focus on sustained work with lower compute, cutting usage costs for businesses. 
• Voice AI becomes a serious channel 
• Anthropic commits to subscriptions only. OpenAI begins testing ads. Incentives now matter. 

Hosted on Acast. See acast.com/privacy for more information.

Microsoft CEO Satya Nadella has said that business applications will eventually “collapse into agents”. In this episode of Future Bytes, Magnus Oxenwaldt speaks with Fredrik Sætre, who owns a Global Black Belt for Microsoft Autonomous AI ERP, about what that shift really means. Not in theory, but in practice. Especially for ERP. 

Highlights from the episode: 

• ERP is not disappearing. The interface is. 
• Agents sit on top of deterministic financial systems. 
• ERP is harder than CRM. Mistakes affect money, tax, and compliance. 
• The real constraint is adoption, not technology. 

Hosted on Acast. See acast.com/privacy for more information.

Earnings from Meta, Microsoft and Apple revealed a shift in AI strategy. Markets rewarded clear returns and questioned vague impact. In this episode, host Magnus Oxenwaldt explains why AI investment now lives or dies on proof.

Highlights for week 6:

• Markets now demand clear AI returns, not long-term narratives.
• Meta showed a direct revenue impact from AI. Others struggled to do the same.
• IBM’s AI growth is driven mainly by implementation, not models or tools.
• Many AI “agents” lack real autonomy, increasing the risk of failed projects.

Hosted on Acast. See acast.com/privacy for more information.

AI moved from promise to proof. At Davos, leaders stopped counting pilots and started demanding impact. Budgets are holding. Returns are not. In this episode, Magnus Oxenwaldt breaks down what that shift means for AI roadmaps in 2026.

Davos signalled a shift, and other key take-aways:

• AI budgets are stable. Measurable impact is not. The gap is now the core challenge.
• Centralised AI centres of excellence deliver significantly higher returns than ad hoc deployments.
• The AI tooling market is fragmenting. Best-of-breed matters more than vendor loyalty.
• Vendor incentives are changing. Governance models need to keep up.

Hosted on Acast. See acast.com/privacy for more information.

Welcome to this week's AI News and the return of our weekly AI briefings for 2026. Listen in to hear host Magnus Oxenwaldt reflections about Apple's decision to use Google's Gemini instead of building its own models, and what it means for your AI strategy.

Highlights: 

• Apple licenses AI models instead of building them 
• Google’s Gemini gains unprecedented global scale 
• AI models move toward commodity status 
• Integration and experience become the defensible layer 

Hosted on Acast. See acast.com/privacy for more information.