“This week our co-CEO and co-founder Jacomo Corbo had the honour to present the keynote address at the Launch Event for #BridgeAI hosted by Innovate UK and to be able to share a few reflections around how #ai system design and adoption patterns are changing - and some of the hard things that aren’t. We are grateful to the UKRI- Industrial Decarbonisation Challenge (IDC) and The Alan Turing Institute for the invitation.
While the frontier of AI capabilities has never advanced more swiftly, transformative digital solutions have also never been more accessible - both to enterprise organizations and to product companies building vertical platforms. ‘How’ AI capabilities are accessed has radically changed in virtue of major advances around infrastructure and platform tooling on which AI systems are built and shipped:
- Productized AI infrastructure has helped simplify and standardize system development and deployment practices -and vastly reduced the need for organizations to build bespoke infra.
- Better platform tooling has moved end-to-end development dramatically up the abstraction ladder (SW2.0), shifting the emphasis in AI development away from deep data science in favour of software engineering tasks focused on configuration and integration and subject matter experts focused on preparing data, improving labels, designing controlled experiments (#datacentricai).
- The shift to AI foundation models, natural language prompt systems and co-pilot tooling (e.g. GPT plug-ins and AutoGPTs) (SW3.0) only reinforces the trend towards solution implementation paths accessible to non-technical users.
The upshot
- Enabled by the right tools, ‘very’ small teams of SMEs and engineers can have outsized impact in building and shipping operational AI capabilities.
- For many AI capabilities, competitive advantage is shifting inexorably away from historical data assets to an organization’s effectiveness at running controlled experiments and building feedback loops.
- Insofar as AI capabilities enable work to be done differently, embedding the benefits requires a commitment to continuous #learning. Fostering a culture receptive to new ways of working remains a leadership challenge. And as the technical barriers to AI adoption fall away, #leadership takes primacy.”