> Markdown version of [/videos/606-when-worlds-collide-how-will-generative-ai-change-the-way-we-design-and-build-software](https://www.wearedevelopers.com/videos/606-when-worlds-collide-how-will-generative-ai-change-the-way-we-design-and-build-software). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# When worlds collide: How will generative AI change the way we design and build software

Consumer chatbots hallucinate without consequence, but enterprise AI demands verifiable facts. Discover how air-gapped generative models and specialized hardware are securely reshaping the future of software development.

- **Speakers:** [Jonas Andrulis](https://www.wearedevelopers.com/@jonas-andrulis), [Mike Butcher](https://www.wearedevelopers.com/@mike-butcher)
- **Event:** World Congress 2023
- **Published:** August 11, 2023
- **Duration:** 44:00
- **URL:** https://www.wearedevelopers.com/videos/606-when-worlds-collide-how-will-generative-ai-change-the-way-we-design-and-build-software

## Summary

As the generative AI boom continues, the industry is splitting between consumer-focused chatbots and highly secure, enterprise-grade solutions. Alef Alpha focuses exclusively on the latter, targeting regulated sectors like government, legal, and healthcare where data sovereignty is paramount. Unlike B2C models that occasionally hallucinate without consequence, enterprise AI requires a rigorous focus on deployment independence. By offering a full-stack system that can operate entirely on-premise and in air-gapped environments, organizations can harness large language models without risking intellectual property leakage or unauthorized external data calls.<br><br>Achieving enterprise trust requires moving beyond basic retrieval embeddings. By manipulating the attention mechanism, developers can trace the flow of factual information down to individual sentence tokens, providing verifiable positive and negative sources that allow human professionals to confidently take responsibility for AI outputs. On the hardware side, scaling dense transformers on traditional GPUs results in inefficient, quadratic compute costs. The next generation of AI architecture relies on modularity, local learning, and conditional sparsity. Utilizing specialized hardware like Graphcore IPUs—which function with a multiple-instruction, multiple-data (MIMD) architecture—enables the complex control flows necessary to mimic the efficiency of biological brains.<br><br>This shift in architecture directly impacts how generative AI will interface with software development. Because software relies on rigid symbolic structures, directly parsing stochastic text outputs from language models remains inherently flawed. Bridging this gap requires new integration paradigms. Beyond technical hurdles, the geopolitical landscape heavily influences AI adoption. Heavy regulation in regions like the EU risks diverting crucial resources away from innovation and toward compliance. Ultimately, the most pressing threat from AI is not existential human extinction, but rather an unprecedentedly fast industrial revolution that will completely reshape knowledge work and accelerate the divide between labor and intellectual property in liberal democracies.

**Keywords:** enterprise generative ai, data sovereignty compliance, air-gapped llm deployment, attention mechanism explainability, retrieval embeddings, dense transformer scaling limitations, conditional sparsity architecture, Graphcore IPU, mimd hardware architecture, symbolic software integration, stochastic text parsing, eu ai regulation, knowledge work disruption, ai intellectual property divide, on-premise foundational models

## Chapters

1. **Transitioning from consumer projects to sovereign enterprise AI** (10:14) — Bypassing the crowded consumer market for highly regulated industries allows for a specialized focus on security and trust.
1. **Tracing knowledge flow through token attention for deep explainability** (13:46) — Manipulating backend attention mechanisms enables users to map positive and negative factual sources directly to individual tokens.
1. **Deploying air-gapped full stack models for critical enterprise IP** (16:11) — Running end-to-end proprietary stacks natively on-premise guarantees complete control and prevents sensitive data leaks.
1. **Securing strategic investments and scaling decentralized engineering talent** (18:26) — Leveraging decentralized university clusters and independent technology stacks attracts top developers efficiently.
1. **Balancing European AI regulation with the speed of innovation** (22:41) — Heavy compliance requirements threaten to divert critical resources away from research in the global technology race.
1. **Open sourcing model architecture to ensure reproducible AI research** (24:48) — Sharing models via permissible licenses encourages community transparency rather than hiding capabilities behind closed product APIs.
1. **Overcoming dense transformer limits using conditional sparsity and IPUs** (27:17) — Utilizing specialized chips for modular local learning drastically reduces the compute requirements of massive current models.
1. **Bridging large language text output with symbolic software structures** (30:06) — Translating conversational prompts into executable code demands robust integration mechanisms beyond simplistic string parsing.
1. **Preparing knowledge workers for the rapid shift in labor** (32:05) — The accelerating automation of cognitive tasks presents a far greater challenge to society than hypothetical existential extinction.

## Related Moments

- [Embedding generative AI in enterprise software platforms](https://www.wearedevelopers.com/videos/916-beyond-the-hype-real-world-ai-strategies-panel) (from "Beyond the Hype: Real-World AI Strategies Panel")
- [Balancing AI regulation with technological innovation in human resources](https://www.wearedevelopers.com/videos/1356-from-learning-to-leading-why-hr-needs-a-chatgpt-license) (from "From Learning to Leading: Why HR Needs a ChatGPT License")
- [Scaling generative AI use cases across large enterprises](https://www.wearedevelopers.com/videos/916-beyond-the-hype-real-world-ai-strategies-panel) (from "Beyond the Hype: Real-World AI Strategies Panel")
- [Crucial lessons for deploying generative AI in enterprises](https://www.wearedevelopers.com/videos/1546-ai-pair-programming-with-github-copilot-at-sap-looking-back-looking-forward) (from "AI Pair Programming with GitHub Copilot at SAP: Looking Back, Looking Forward!")
- [The shift towards generative artificial intelligence in production](https://www.wearedevelopers.com/videos/969-make-it-simple-using-generative-ai-to-accelerate-learning) (from "Make it simple, using generative AI to accelerate learning")
- [Overview of enterprise Java and generative AI](https://www.wearedevelopers.com/videos/1554-java-meets-ai-empowering-spring-developers-to-build-intelligent-apps) (from "Java Meets AI: Empowering Spring Developers to Build Intelligent Apps")

## Related Articles

- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)
- [WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?](https://www.wearedevelopers.com/magazine/469-wwc24-talk-scott-hanselman-ai-superhero-or-supervillain)
- [Navigating the AI Shift](https://www.wearedevelopers.com/magazine/629-navigating-the-ai-shift)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)

## Related Jobs

- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Partner Sales Director - AI Alliances - Model Providers](https://www.wearedevelopers.com/jobs/48429-partner-sales-director-ai-alliances-model-providers) at **Dynatrace**
- [Senior AI Developer](https://www.wearedevelopers.com/jobs/ext/2836034-senior-ai-developer) at **PwC**
- [Senior AI/ML Engineer](https://www.wearedevelopers.com/jobs/48352-senior-ai-ml-engineer) at **PagerDuty**
- [Head of Agentic AI / Lead AI Engineer](https://www.wearedevelopers.com/jobs/48488-head-of-agentic-ai-lead-ai-engineer) at **1st solution consulting gmbh**
- [Principal Software Engineer, AI Compute Platform](https://www.wearedevelopers.com/jobs/ext/2847710-principal-software-engineer-ai-compute-platform) at **ARM**