World Congress 2023 • Aug 11, 2023

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

Jonas Andrulis , Mike Butcher

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.

Pause
Mute Enter Fullscreen
#1 about 4 min

Transitioning from consumer projects to sovereign enterprise AI

Bypassing the crowded consumer market for highly regulated industries allows for a specialized focus on security and trust.

#2 about 3 min

Tracing knowledge flow through token attention for deep explainability

Manipulating backend attention mechanisms enables users to map positive and negative factual sources directly to individual tokens.

#3 about 3 min

Deploying air-gapped full stack models for critical enterprise IP

Running end-to-end proprietary stacks natively on-premise guarantees complete control and prevents sensitive data leaks.

#4 about 5 min

Securing strategic investments and scaling decentralized engineering talent

Leveraging decentralized university clusters and independent technology stacks attracts top developers efficiently.

#5 about 3 min

Balancing European AI regulation with the speed of innovation

Heavy compliance requirements threaten to divert critical resources away from research in the global technology race.

#6 about 3 min

Open sourcing model architecture to ensure reproducible AI research

Sharing models via permissible licenses encourages community transparency rather than hiding capabilities behind closed product APIs.

#7 about 3 min

Overcoming dense transformer limits using conditional sparsity and IPUs

Utilizing specialized chips for modular local learning drastically reduces the compute requirements of massive current models.

#8 about 2 min

Bridging large language text output with symbolic software structures

Translating conversational prompts into executable code demands robust integration mechanisms beyond simplistic string parsing.

#9 about 3 min

Preparing knowledge workers for the rapid shift in labor

The accelerating automation of cognitive tasks presents a far greater challenge to society than hypothetical existential extinction.

Matching moments

3:13 min

Embedding generative AI in enterprise software platforms

Mike Butcher Mike Butcher +3 · World Congress 2024

3:17 min

Balancing AI regulation with technological innovation in human resources

Rudi Bauer Rudi Bauer +1 · Cappuccino with HR

2:52 min

Scaling generative AI use cases across large enterprises

Mike Butcher Mike Butcher +3 · World Congress 2024

3:33 min

Crucial lessons for deploying generative AI in enterprises

Alexander Trusheim Alexander Trusheim +1 · World Congress 2025

1:38 min

The shift towards generative artificial intelligence in production

Duan Lightfoot Duan Lightfoot · World Congress 2024

54 sec

Overview of enterprise Java and generative AI

Timo Salm Timo Salm · World Congress 2025