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

What if you could verify every claim an AI makes? Discover how explainable, sovereign AI is building trust in the systems that matter most.

#1about 2 minutes

Introducing Aleph Alpha and its European AI vision

The founder of Aleph Alpha discusses the company's journey and its positioning as a European player in the generative AI space.

#2about 1 minute

Focusing AI on high-stakes enterprise use cases

Aleph Alpha targets critical B2B sectors like government and finance where data sovereignty and reliability are more important than consumer applications.

#3about 2 minutes

Tracing information flow for true AI explainability

A proprietary manipulation of the attention mechanism allows tracing factual claims back to specific sources, including identifying contradictory information.

#4about 4 minutes

Offering a full-stack AI solution for on-premise deployment

The platform is a complete, end-to-end stack that can be deployed in isolated, on-premise environments to ensure customers have full control and data privacy.

#5about 2 minutes

Building an AI talent hub outside major tech capitals

Locating in Heidelberg provides access to strong engineering talent from nearby universities and a unique company mission that attracts inbound tech applicants.

#6about 2 minutes

Arguing against heavy EU regulation on AI

Heavy regulation could divert critical resources from innovation to compliance, slowing Europe's ability to compete and secure its digital sovereignty.

#7about 2 minutes

Balancing open source contributions with commercial IP

Contributing to the community through open source is vital for reproducible research, but it must be balanced with protecting the IP needed for monetization.

#8about 3 minutes

Partnering on hardware for next-generation AI models

The partnership with Graphcore aims to overcome the limitations of GPUs for inefficient dense transformers by developing hardware for more modular, brain-like architectures.

#9about 2 minutes

Bridging the gap between language models and software

The next frontier in software development is creating a robust interface between stochastic large language models and structured, symbolic software systems.

#10about 12 minutes

Prioritizing societal challenges over AI doomsday scenarios

The more pressing danger from AI is not existential risk but the rapid disruption to knowledge work and the challenge of adapting our societal values and institutions.