Budoucnost patří AGI a robotům: Jaroslav Beck o životě po prodeji Beat Saber

Business Summary

Current Ventures and Strategy

• BottleCap AI: Beck’s main focus is a European AI research center he is building with Tomáš Mikolov and David Harel. The team is researching model architecture and efficiency, with the goal of developing its own technology and products in-house—not merely wrapping existing models in another chatbot.

  • Beck said the team had developed an algorithm that could make model training about 50% faster, potentially reducing training costs substantially. He presented this as an early result, not a finalized commercial outcome.
  • At the time of the interview, the team was assembling and recruiting researchers. Beck said it hoped to release its first products early the following year, then publish at a faster cadence.
  • The NoCap Test asks candidates to improve a baseline model’s performance or training efficiency.

• Cans: Beck describes this beverage company as a difficult but potentially scalable consumer-goods business. Its product goes against a market dominated by sugar and artificial sweeteners.

  • Unlike software, the business faces low margins, manufacturing and import costs, staffing and marketing expenses, and the challenge of winning supermarket shelf space.
  • Beck said traction was increasing month by month and projected that Cans could become profitable in the Czech Republic the following year. He characterized this as a fast timeline for this type of business.
  • He estimated that entering a new market can require roughly six months of investment before a brand gains meaningful adoption.
  • The intended sequence is to establish a self-sustaining business in the Czech Republic and Europe, including the UK, before taking on the US market. He sees success in the US as a large upside, but not a prerequisite for the business to survive.

• Beat Saber and the founder’s role: Beck describes himself as especially effective at taking projects from zero to 100: recognizing potential, energizing a team and pushing a product toward broad impact. He is less suited to the later-stage work of scaling an established operation from 100 to 200, where other leaders may be stronger. He said he knew when Beat Saber no longer needed his input and could stop improving it indefinitely.

Operating Principles and Playbooks

• Start with a real problem and useful impact. Beck’s project filter combines solving a real problem with the potential to help many people and become a viable, scalable business. Impact and commercial viability should reinforce each other.
• Build a complementary team. He argues that no founder is strong at everything; an effective company brings together people who cover one another’s weaknesses. In hiring, he looks for:
  • Strong drive and enthusiasm for the work.
  • People he can learn from and who have expertise to contribute.
  • Evidence that their energy can be contagious to the rest of the team.
• Use simple, first-principles questions. Ask basic questions until the underlying idea is clear. If a theory or plan cannot withstand a simple question, either the team does not understand it or the problem has been framed poorly.
• Protect relationships and keep attention on the goal. Beck prefers direct discussion and long-term problem-solving over conflict for its own sake. He tries not to react impulsively or let ego derail the objective. Escalate when necessary; otherwise, resolve the issue and move on.
• Be transparent rather than performative in marketing. Beck has not used a PR agency to run his social presence. He favors explaining decisions in his own voice—often in longer conversations—over manufactured authenticity or lifestyle promotion. A product should be the substance, not an excuse to sell a founder’s image.
• Treat the founder role as a fit, not a universal ideal. Founding is difficult and risky. Beck’s venture investments are based on the belief that a company addresses a genuine problem, not on confidence that he can predict whether it will succeed. He says most startups he backs will fail and does not regard himself as a reliably skilled investor.

Time Management and Execution

• Beck’s “JB Real Time” approach is to handle a message immediately when it can be resolved quickly. His reasoning is that keeping a task in mind consumes attention and increases the chance of forgetting it.
• His email routine distinguishes between quick actions and work requiring thought: he answers the former immediately and batches more involved replies for later. He said receiving more than 200 emails a day can become irritating, and that he aims to maintain a zero inbox.
• His calendar is his central task system. He records ideas and work there, often in 10-minute blocks, and moves unfinished tasks forward until he completes or deliberately removes them.
• He acknowledged that immediate replies and frequent context switching will not suit everyone. He is also experimenting with a private, offline AI trained on his own writing and preferences to draft routine email responses, while warning against casually sending all personal correspondence to online AI services.

Business Examples and Operational Choices

• Beat Saber: Beck joined after seeing an early demo, recognized its potential and helped push it toward global reach. The example illustrates his “zero-to-100” strength: act on promising opportunities quickly, then recognize when further founder involvement is no longer essential.
• Cans: Rather than optimize only for yield, Beck says the team chose properties it could renovate into attractive places. He prefers a smaller number of larger, high-quality apartments—for example, six three-bedroom apartments rather than 50 small units—to reduce operational complexity and create a better product for residents.
• Hiring: BottleCap’s test makes the work itself the assessment: candidates are asked to improve a model-training baseline rather than being evaluated solely on conventional credentials.
• Learning and career decisions: Beck recommends trying a range of things early, when personal commitments are lower, to discover what is genuinely motivating. He sees a startup as a way to learn through intense, practical work; university can offer valuable social connections and education. His view is conditional: a rare startup opportunity may be worth taking because formal education can often be resumed later.

Metrics and Targets Mentioned

• BottleCap: A reported ~50% improvement in training speed from an early algorithm; first products targeted for early the following year.
• Cans: Traction reportedly rising monthly; target of Czech-market profitability the following year; around six months of market investment may be needed before brand adoption.
• Email load: Over 200 messages per day can become burdensome; calendar tasks are often scheduled in 10-minute intervals.
• Investing: Discussed as personal context rather than operating advice, Beck cited a portfolio worth about 15× its earlier value and roughly 50% annualized performance over four years, while emphasizing that he is an amateur and that luck played a major role. He said he keeps investing from becoming his primary focus because building companies matters more to him.
• Venture portfolio: Beck said he had invested more than 100 million in startups, treating the amount as money that could be lost rather than assuming it would contribute to his personal net worth.

Technology Outlook

Beck expects autonomous cars to become more common, subject to regulation, and believes AI will displace some existing work while creating roles that do not yet exist. He expects more capable robots to take on practical tasks, but distinguishes task-specific robots from fully autonomous household helpers. In his view, the combination of capable AI and a physical, general-purpose robot is what will make “AGI” feel tangible to many people.

Presenters and Sources Mentioned

• Presenters: Jaroslav Beck (guest) and Kirill Juran (interviewer/host).
• People and sources referenced: Tomáš Mikolov, David Harel, Honza Ilavský, Vlad Hrnčiar, Walter Isaacson’s biography of Steve Jobs, Elon Musk, Mark Zuckerberg, Hans Zimmer, Michal Ptáček’s interview with David Semerád, Forbes, CNBC and the BottleCap AI NoCap Test.