Slide contents

1. 1. Software development in 2030

   Testable hypotheses instead of a linear forecast

2. 2. Alexander Polomodov

   Technical Director & Fellow, large fintech

   Architecture and engineering practices in fintech

   AI-native development and Platform Engineering

   Code of Leadership podcast, @book_cube

3. 3. The future resists linear extrapolation

   Test every hypothesis with four questions

   Which signal can we observe now?

   What must change by 2030?

   What could break the scenario?

   Which metric would falsify it?

4. 4. 01. Signal: agents become work units

   But an autonomous task is not autonomous development

5. 5. From one agent to an agent system

   The question mark matters more than the arrow: stage three is unproven

6. 6. Protocols lower integration cost; they do not create trust

   Connecting an agent is easier than delegating safely

   MCP connects models to context and tools

   A2A describes interaction between agents

   Permissions, audit and accountability stay local

   Interoperability ≠ autonomy

7. 7. Adoption cannot prove that delivery has improved

   Tool popularity is input, not outcome

   No provenance for '70% use it daily'

   Generated code may increase review volume

   Local speed does not guarantee system gains

   We need delivery, quality and outcome metrics

8. 8. 02. Research offers signals, not dates

   AlphaEvolve, WORKBank, agent economies and SE 3.0

9. 9. A verifiable evaluator turns generation into search

   Strong results appear where a solution can be scored automatically

10. 10. Algorithm success does not prove full-SDLC autonomy

    DeepMind reports real results and a clear applicability condition

    Data-centre, chip-design and AI-training optimisation

    New matrix-multiplication algorithms

    Evaluation must be objective and automated

    Ambiguous product goals are not scored this way

11. 11. Automation capability and worker desire: separate axes

    A model being able to do a task does not mean it should be deployed

12. 12. Agent economies differ by origin and permeability

    The path the authors consider most likely is also the riskiest

13. 13. 03. Scenario: intent becomes the interface

    A human states the goal; a machine searches the solution space

14. 14. AI-native shifts focus from code to intent

    This is a research programme, not a description of an era already here

15. 15. Measure durable goals before choosing tools

    Thirty goals connect sentiment and behavioural signals

16. 16. A large company needs a runtime for safe delegation

    The model is only one layer of an agent platform

    VCS, CI/CD, observability and a service catalogue

    Context, permissions and policy enforcement

    Isolated execution and reproducibility

    A human approves high-risk changes

17. 17. Agents span the SDLC without transferring accountability

    Every phase needs a testable goal and an explicit decision owner

18. 18. Agent-written code requires scalable verification

    The constraint is proving quality, not generating code

    Evaluators must test functionality and risk

    Review cost belongs in total task cost

    Rising rework falsifies the productivity promise

    Safe rollback limits the blast radius

19. 19. 04. Metrics connect agents to outcomes

    Acceptance rate helps, but does not show whether the system improved

20. 20. No single metric survives agentic development

    Combine flow, experience, agent work and user outcomes

21. 21. A pilot needs a stop condition before it starts

    Otherwise local acceleration is easy to mistake for impact

    A baseline on comparable tasks

    Task success and human intervention

    Total cost: compute + review + rework

    Escaped defects and delivery outcome

22. 22. One signal can lead to different futures

    Watch which scenario the data supports, not how much code is generated

23. 23. By 2030, verifiable systems beat autonomous ones

    What evidence must confirm or falsify

    Intent becomes more important than typing code

    Evaluators become part of development architecture

    The platform defines the agent autonomy boundary

    Outcomes and rework outrank generation volume

24. 24. Observations are separated from the author's scenarios

    Full links are in the speaker notes

    DeepMind: AlphaEvolve · Virtual Agent Economies

    Stanford: Future of Work with AI Agents

    SE 3.0: AI-native Software Engineering

    Google Research · DORA: developer goals and metrics

25. 25. Thank you!

    polomodov.tech

    All slides and links — in the Telegram channel

    Alexander Polomodov, Technical Director & Fellow, large fintech

    @book_cube