Essays on Bruno's Journal

GenAI: Competing with Open Source and Proprietary Data

Tue, 17 Sep 2024 00:00:00 +0000

A couple of years ago, I started seeing young startups raising early-stage venture capital funding and diving into the GenAI revolution. Investors are betting that GenAI will open the door to a new Industrial Revolution, where GenAI transcends from high-tech to becoming ubiquitous across agricultural, industrial, and service sectors.

What is becoming obvious is that this revolution carries a significant increase in capital investments, especially when compared to previous cycles. In a way, what’s happening now is not dissimilar to what we saw in the late 1990s, when thousands of servers were being racked up in dedicated colocation facilities, financed with VC money to create consumer experiences pursuing a blitz economics paradigm. Eventually, the boom was followed by the 2000 burst, and the lack of profits and huge capital expenditure led to only a few companies—those that had been very conservative in their capital spending and profitability—surviving and becoming the tech giants of the next decade, such as Yahoo, Amazon, and Google.

Rethinking R&D Structures for GenAI Startups

Tue, 17 Sep 2024 00:00:00 +0000

The GenAI revolution is rapidly transforming industries worldwide, and the automotive autonomous vehicle (AV) sector is at the forefront of this seismic shift. Startups like ours, leveraging open-source technologies and proprietary driving datasets, are compelled to reevaluate traditional research and development (R&D) structures to stay competitive.

The disparity in spending power between tech giants and smaller companies has never been more pronounced. While corporations invest billions into artificial intelligence advancements, startups must innovate organizationally to bridge the gap without matching such colossal expenditures. Simultaneously, academia and grassroots innovators grapple with challenges, striving to remain relevant in an environment dominated by deep pockets and rapid technological progress.

The Nexar Network Econometrics

Sat, 17 Jun 2023 00:00:00 +0000

This blueprint helps understand why we work on certain things, and why we work on it now. But in addition to this it’s important to also be able to justify the priorities of what we work on. More precisely what’s the allocation of relative effort and time we should put in one feature vs another.

A highly pretentious model to understand product prioritisation

What we do at Nexar falls into three buckets:

Organisational Structures & Generative AI Applications

Sat, 03 Jun 2023 00:00:00 +0000

Deciding which organisational structure is most efficient for pushing forward research & development initiatives, taking into account the interaction between research, engineering, and product teams, has long been a big topic of discussion as it is paramount as a precursor of success. Organisational behaviour and structures are deemed to change in the era of Generative AI. Our success in applying Generative AI technology to Nexar will be large conditioned in our ability implement the right organisational structure to optimises productivity and velocity.

The Formidable Power of Code

Thu, 19 Jan 2023 00:00:00 +0000

After writing last night the piece on engineering culture, Why Nothing Works: Kings and Priests, I also wrote about it in tangentially in Heroes and Generals, but except a few at Nexar, nobody has had a chance to interact with how I think of code, how I write code, and why I think code is the one thing that allows measuring whether an organisation is working, or not. My apparent failure in the first few months in Sunnyvale, how did I succeed? The answer is code.

Tail for Tails

Mon, 24 Oct 2022 00:00:00 +0000

Software innovation is usually seen as a improvement on existing processes. However, the critical software innovation is not incremental, it’s built like a ladder. Google is famous as an organization for shooting for the moon, and not investing resources on a project if it won’t deliver at least one order of magnitude improvements. These type of changes do not result in changes to existing processes, but rather the creation of new processes that altogether replace less efficient ones, disrupting the space of the incumbents.

The Ecosystem Keystone

Mon, 24 Oct 2022 00:00:00 +0000

An ecosystem grows a network faster than any company can grow it alone — for Nexar, at least 10x faster, through rich multi-lateral B2B relationships and partnerships. Ecosystem deals expose positive externalities: the sum of the parts is greater than the parts alone. The question is what and why the ecosystem can work, which then leads to how to make it work.

In high-tech, healthy ecosystems are characterized by a “keystone”. As partners develop 1:many relationship within the ecosystem, there is always an invariant, which is the keystone. In Intel’s ecosystem, the keystone is the CPU. There are then value-added products on top of it, from components and drivers, operating system(s), application software, computers, etc. Each player develops its own partnerships, e.g. HP with Microsoft for OS, HP with Samsung for SSD, etc. but it must always revolve around the keystone, the CPU.

Leadership through Storytelling

Sat, 16 Apr 2022 00:00:00 +0000

A few months after becoming head of personalisation and advertising at Yahoo, I had to get onstage at the sales conference in Las Vegas in front of over 2,000 sales employees, all hungry for me to bring to them things to sell. I was supposed to tell them about the implications and future of merging personalisation and advertising. During the rehearsal with my public speaking coach, I started narrating the presentation I had prepared by talking about the ideas I wanted to pass on to the audience, only to be quickly interrupted by the coach in order to tell me that it was boring, aseptic, and that these sales people would forget about me and all the work my team was doing within minutes. Instead she told me to make it personal, to make the presentation about my personal story. So instead of telling them about personalisation and advertising, I told them about how I had to put my twins to sleep every night through storytelling, customising the stories to each of them, but at the same time blending them since they slept in the same room. Even though this was a very poor attempt at storytelling, I then learnt that I had to become better at storytelling, and even more, I touched the exciting power of personal storytelling. Over time I learnt to tell stories about things that had happened to me whether at work or in my private life in a coherent way to convey my guidance and leadership.

Technical Leadership Skills

Sat, 16 Apr 2022 00:00:00 +0000

There are business books by the dozen written about leadership, but we keep struggling to clearly define technical leadership. Even though we seem to intuitively acknowledge that leadership is a thing, we can’t clearly qualify it, nor quantify it. Technical leadership applies equally to people managers and individual contributors. The rubrics that define a technical leader don’t change because one writes code and the other one does not. The degree in which that rubric is exposed does, but not the trait itself.

Why Nothing Works: Kings and Priests

Sun, 10 Apr 2022 00:00:00 +0000

After 2 years working for Yahoo in London, Yahoo outsourced search to Microsoft Bing. I moved to the US to take on the role of chief architect for the Yahoo Platform, and I was faced with the problem of getting things done. Now that might seem like an obvious thing, “getting things done”, but it was not. After two months in Sunnyvale, I still could not figure out how to get any results. I was failing. I had started at Yahoo working in London responsible for Yahoo search outside of the US, as an architect, an individual contributor with no reports. Although I had no direct authority, International engineering was a relatively small group of about ~1000 engineers that had a highly strict and disciplined engineering culture that John Linwood, my boss, had brought from Microsoft. Inserting myself into that structured culture was easy. I found where the reviews happened, where to position myself so that the right decisions were made before it was too late. But Yahoo in the US lacked that engineering discipline, or at least I could not explain it. And I did not know how to navigate it. All I could do as an architect was exercise influence through “gravitas” as I had no reports. What had worked in London wasn’t working in the US.

Heroes or Generals?

Wed, 30 Mar 2022 00:00:00 +0000

There is one gap that I identified for which I am ultimately responsible as CTO, culture, that I would like to cover and help us adjust. It relates to what we expect out of managers at Nexar.

Most “traditional” hi-tech career management approaches follow the up-or-out career model for the management track. Sometimes it was explicit (you were actually managed out if you did not promote), and sometimes it was passive (you stalled and received zero salary increases). The rationale of the model was to pick the best managers out of the system. “Best” in this model means those that are able to:

Now is the Time for Our Developer Platform

Mon, 28 Mar 2022 00:00:00 +0000

Over the years we have evolved a dream, a vision for a safer and more intelligent future, into a company impacting the lives of hundreds of thousands of people who have or are driving with Nexar. We have also built systems to make this dream a reality. We started a team creating our platform, initially simply a team developing the server-side functionality required by the iOS and Android applications. Then we decided to shift gears and move into what we then called the “data platform” and make our platform the foundation for many AI-verticals we would go on and build based on the data that the Nexar network collects.