At Tech Race Summit 2026 in Warsaw, Coinspaid Dev Executive Leader Alexey Tulia urged technology leaders to build organisations that can absorb wrong bets. His reasoning starts from a simple observation: AI is about to gain the power to act directly inside company infrastructure.
As noted by BlockchainReporter, Tulia shared his views on the AI Impact in Engineering panel, which looked at how the roles of engineers and CTOs are evolving. For him, any technology budget should begin with a concrete organisational need. Instead of chasing each new AI product, he recommends putting money into capabilities that allow a company to adopt new technology without endangering its operations. These include robust APIs and trustworthy data, backed by automated testing, observability, security and an architecture that can adapt as requirements shift.
Flexibility also depends on time. Tulia pointed out that a roadmap filled to the last hour leaves engineers no chance to evaluate an emerging tool or change direction when priorities move. Teams therefore need deliberate capacity for experiments. The same thinking applies to architecture and to limiting dependence on a single vendor. Such work rarely shows up in short-term revenue figures, but it lets a company swap providers or rebuild parts of a system once its original assumptions stop being true. Tulia explained that he has no need to foresee the future perfectly. What he needs is for errors to be inexpensive to correct.
That preparation becomes critical as AI moves beyond its current supporting tasks, such as drafting documents and analysing data. The next step Tulia described is connecting agents to live systems, including sensitive information and deployment pipelines, where they can take real actions. Consider an agent able to prepare a code change and release it to production. It is not yet obvious whether it should do so without human approval, or who would be held responsible if the release caused a failure. Before granting that kind of access, Tulia said, an organisation should have the following in place:
- permission controls that limit what the agent can reach;
- audit logs that record every action it takes;
- a dependable way to stop the agent immediately;
- a tested process for recovering from a failed deployment.
The principle behind this checklist is that accountability must keep pace with the authority given to machines. More autonomy in production is acceptable only when authority is clearly defined and a human remains responsible for the outcome.
AI is changing the engineer’s role as well. With coding and prototyping moving faster, Tulia sees an opportunity for engineers to spend more effort understanding the business problem and to stay involved as their work runs in production. Leaders can encourage this by giving teams the business context behind a task and a clear description of the expected result. Productivity is then measured by correctness, maintainability, security and operational performance, while the sheer volume of code loses its importance as a metric.
Looking further ahead, Tulia expects that by 2029 smaller engineering teams will manage broader areas of responsibility and that AI will write most production code. Verification and technical judgment will grow in value as a result. The CTO position will still demand deep technical knowledge together with an understanding of the business, especially as simpler software creation brings more vendors and AI-generated systems into each company. In Tulia’s view, technical judgment will matter even more than it does today. His immediate recommendation for engineering leaders is to agree on safeguards and ownership before any AI agent is allowed near critical production systems. Tulia leads at Coinspaid Dev, an independently owned software engineering company specialising in blockchain infrastructure, with more than 120 engineers, over 11 years in the industry and experience operating across more than 20 blockchain networks.