Notes from Brazil at Silicon Valley (BSV) 2023
Last week, I went to the fifth edition of the Brazil at Silicon Valley (BSV) conference. In general, I felt that the vibe of the event was optimistic, which was a surprise for me given the current market conditions. Some of the positive ideas shared were how Artificial Intelligence (AI) should be used as a copilot for humans while keeping us at the center of work, the case for a social media platform that promotes happiness and meaningful connections instead of what is more profitable, and how innovative tech solutions can help solve the ongoing climate crisis.
There were also, of course, many challenges discussed by the panelists and attendants. Among them, the most recurring ones were the critical role that regulators have to play globally to prevent bad actors in this new AI era and the challenging funding environment for both founders and VCs that will and already are resulting in down rounds for startups. Below, I share the notes I took during some of the panels that I attended.
Angela Strange, GP at Andreessen Horowitz, and Hugo Barra, CEO at Detect
Panel name: Why is now a great time to build in Brazil?
About the speakers: Angela is a GP at Andreessen Horowitz where she focuses on fintech and insurance investments. Prior to a16z, she was a Product Manager at Google where she launched and grew Chrome for Android and Chrome for iOS. Hugo is currently the CEO of Detect. Among many other roles, he has previously worked with Angela as Vice President of Android Product Management at Google.
Notes:
Angela was the one who coined the idea that “every company will be a fintech company”. In summary, she believes that any company can and should provide financial services to better, attract, retain, and monetize its users, something that can be applied to both B2B and B2C businesses.
An example would be how Uber now serves as a sort of bank for its drivers since they can earn, save, and spend money directly from Uber’s wallet.
A LatAm example would be Fudo, which started as a management software for restaurants that also had a financing lens from the get-go.
Angela is impressed by how some regulatory bodies in Brazil run like a private tech company. As an example, PIX’s adoption curve is similar if not better than other great private fintech products. This was only possible because it had a structured product roadmap and business plan.
Two other initiatives from the regulatory bodies that may benefit new fintechs are Open Finance and how much easier it became for newcomers to get licenses for some financial services.
Brazil has tons of problems, there are so many industries we can bring tech to. She thinks the entire financial infrastructure in the region could be rethought with modern APIs that would allow for way more efficient financial services and better user experience.
Talking about AI, Angela mentions that humans tend to overestimate what can happen in 1 year but underestimate what could happen in 10 years.
AI can be key for increasing financial inclusion. Fintechs can leverage this technology to lower their cost to serve, which could allow them to serve customers that weren’t profitable for them before, such as lower income and more senior people.
Brazilian founders are great at spotting raw talent and developing them.
Brazilian founders are not great at selling their vision and accomplishments to investors (differently than Bay Area ones).
Alexandr Wang, CEO of Scale AI, and Peter Norvig, Researcher at Google
Panel name: Democratizing Artificial Intelligence
About the speakers: Alexandr is the CEO and Founder of Scale AI, a company that is valued at over $7B and provides a data platform that enables high-quality training and validation data for AI applications. Peter is a Researcher at Google and was formerly a lecturer at Stanford, Head of Computational Sciences Division at NASA, among many other roles.
Notes:
Data is going to be critical for the realization of AI and data is not the new oil. Unlike oil, data is not fungible. A dataset from Brazil is very different than a dataset from the US for instance and they might train algorithms in totally different ways.
Both agree that we shouldn’t press pause on the development of AI models. In their view, AI organizations are already careful with security and privacy. This is already the number one priority of teams developing these models but regulatory clarity is the best way to prevent bad actors.
We should integrate AI tools to augment and improve how we work. If we bring computing costs down, we could make these tools as widespread as possible so that everyone would become the “manager of a bunch of robots”. In that sense, AI could help to distribute wealth and not make the inequality gap even wider.
Vinod Khosla, Founder of Khosla Ventures, and Julio Vasconcellos, Founder and Managing Partner at Atlantico
Panel name: Technology Trends and the Importance of Bold Leadership
About the speakers: Vinod is a legendary entrepreneur and venture capitalist. He is the Founder of Khosla Ventures, a “venture assistance” firm that helps entrepreneurs working on breakthrough technologies. He previously co-founded Sun Microsystems. Julio is the Founder and Managing Partner at Atlantico, an early-stage, LatAm-focused VC firm. Among several other experiences, he was most notably the CEO and Founder at Peixe Urbano and Brazil Country Manager at Facebook.
Notes:
The entrepreneurial journey is not for everyone. The highs are very high and the lows are very low. People tend to see only the good part.
Founders need to take bold risks and embrace failure to be successful.
Good entrepreneurs are always learning and flexible to adapt to situations. They need to be comfortable with ambiguity but still be driven by a strong vision that they want to make a reality.
Most investors add negative value to their portfolio since they look at the wrong metrics. One thing that they are usually good at is connecting entrepreneurs to people who have gone through similar experiences.
The expense of how AI will change things is still unclear. However, founders should think about how AI could be used as if they were starting their businesses today.
Orkut Buyukkokten, Founder of Orkut & Hello Network
Panel name: Orkut, Social Media, and Mental Health
About the speaker: Orkut is the Founder of Orkut. He is originally Turkish and founded the social media platform while he was working as an engineer at Google when the company allowed any worker to devote 20% of their time to a personal project. He first started building social networks because he wanted to make it easier for people to meet and make more meaningful connections with each other.
Notes:
Social networks’ algorithms are designed to increase engagement and not happiness. Posts that contain fake news or harmful content are usually the ones that get more engagement and thus get promoted by social media companies. This contributes in some way to the rise in depression and other mental health issues in young people today. Orkut believes that there is a better way to build a social media platform in a way that promotes happiness and meaningful connections. That’s what Orkut did and what he is trying to do again now.
Ali Dalloul, VP Open AI & Azure AI Microsoft
Panel name: The Promise of Generative Artificial Intelligence
About the speaker: Ali has spent almost 25 years working at Microsoft and now leads the Azure AI and Azure Open AI customer experience engineering team.
Notes:
Ali believes that humans will always be at the center of work but AI will be a valuable assistant and a catalyst for creativity. A good example of this dynamic is Microsoft’s Copilot, a work copilot that uses large language models (LLMs) and internal company data to assist users in presentations, spreadsheets, written documents, etc.
The AI revolution is similar to the cloud revolution in terms of its impact. Cloud technology brought computing costs dramatically lower, which enabled the quick development of software startups.
If you’re starting an AI business today, you still need to deeply solve a human need. This is still the only way to maintain your defensibility in the long term.
We should embrace AI and not resist it. People need to learn how to “code Chat GPT” or in other words, learn how to ask the right questions and take the most out of the tool.