#12 Privacy in the AI Era
Plus: Cloud benefits from AI, Amazon Q, AI-First product mindset, pros and cons of specialized board directors, AI use in aquaculture, and more
👋 Welcome to "AI Simplified for Leaders," your weekly digest aimed at making sense of artificial intelligence for business leaders, board members, and investors. I invite you to explore the past issues here. If you like this newsletter, please forward it to your friends. I greatly appreciate it.
In this issue, I cover:
AI news: cloud players benefit from AI boom; Amazon Q; OpenAI gpt2-chatbot rumor, and more.
An inside look at how Perplexity builds products with an AI-first mindset
Privacy in the AI era: what leaders need to know
Director’s Corner: should boards add specialized directors?
AI is helping to put healthy seafood on your table
Enjoy.
AI News, Curated for You
1. Cloud infrastructure players are already monetizing from the AI boom
As mentioned here quite a few times, cloud infrastructure players are best positioned to influence business users' AI model selection, pick AI startup winners with their platform and ecosystem support, and benefit from the AI application boom. The performance divergence between big tech with cloud (Microsoft, Amazon, Google) and without (Meta) after recent earnings adds credibility to this hypothesis.
Jamin Ball from Altimeter Capital is known for tracking operational metrics for cloud players. Glancing at the year-over-year growth of cloud revenue growth at Amazon AWS, Microsoft Azure, and Google Cloud, you will see a clear reacceleration of growth since Q1 2023, when ChatGPT 4 came out.
Meanwhile, I am puzzled by the valuation of the so-called AI-dedicated cloud player. CoreWeave, an AI cloud computing startup, recently raised $1.1 billion in a funding round, catapulting its valuation to $19 billion. This substantial valuation marks a near-tripling from its previous $7 billion valuation just a few months prior. Let me know if you can help me understand the investment thesis. CoreWeave certainly made a great pivot from hosting GPUs for crypto mining a few years ago.
2. Amazon Q Business impresses with its simplicity
Amazon launched its Amazon Q Business, its AI business assistant. From what I have seen and tried so far, it is incredibly simple to deploy and manage, with built-in security and governance features. Employees can make informed decisions by accessing data-driven insights and internal information without worrying about data leakage.
I am excited to see that Amazon aims to help non-technical leaders launch their internal AI applications on data insights quickly. This helps to deal with the big headache of finding the right talent to lead AI initiatives and further supports my proposal of the type of AI leaders businesses may need. Read more here about why Forbes’ Steve McDowell believes Amazon Q is a leap forward.
3. Other launch news (and rumor)
OpenAI ChatGPT now has memory across chats. If you enable the option in ‘settings’, it will remember what you tell it in past conversations and act more like a personal assistant.
Anthropic launches its mobile app and first enterprise offering. Many people have switched from ChatGPT to Anthropic for its more business-oriented and succinct style. Could this change with the OpenAI May 9 announcement (not confirmed)?
A mysterious gpt2-chatbot emerged briefly online on April 29 and its performance seemed to surpass GPT4 and Llama3 in some tasks including reasoning. People speculated that it was a version of OpenAI’s GPT 4.5, and Sam Altman’s tweets encouraged such speculations. We will see.
4. OpenAI and Financial Times strike license deal
Under the terms of the deal, OpenAI can train its models with the archive of the Financial Times. Financial Times has many insightful articles looking at financial histories and business models that are timeless - it is great to see this collaboration opens up a legal and responsible way of letting gen AI models tap into the power of monetizable insights in finance.
5. BCG says AI consulting is 20% of business in 2024
The global consulting firm BCG says AI consulting is 20% of its revenue this year and will reach 40% by 2026.
How Does an AI-First Company Build Products?
- A Fascinating Inside Look at Perplexity
Readers of this newsletter would know that I use Perplexity for most of my search needs. (I hope you give it a try too!) The two-year-old company has fewer than 50 employees, tens of millions of users, and a recent valuation of over $1bn.
For Lenny Rachitsky’ recent newsletter, he sat down with Perplexity’s co-founder and head of products, Johnny Ho, to discuss how the company builds products.
My biggest impression is minimalism in product development. Here’s how Lenny summarizes the way Perplexity works:
AI-first: They’ve been asking AI questions about every step of the company-building process, including “How do I launch a product?” Employees are encouraged to ask AI before bothering colleagues.
Organized like slime mold: They optimize for minimizing coordination costs by parallelizing as much of each project as possible.
Small teams: Their typical team is two to three people. Their AI-generated (highly rated) podcast was built and is run by just one person.
Few managers: They hire self-driven ICs and actively avoid hiring people who are strongest at guiding other people’s work.
A prediction for the future: Johnny said, “If I had to guess, technical PMs or engineers with product taste will become the most valuable people at a company over time.”
Many interesting tidbits in the interview will lead you to rethink your long-held assumptions about organizations and workflows, even if you are not a product leader. Here is an example, as described by Johnny:
My goal is to structure teams around minimizing “coordination headwind,” as described by Alex Komoroske in this deck on seeing organizations as slime mold. The rough idea is that coordination costs (caused by uncertainty and disagreements) increase with scale, and adding managers doesn’t improve things. People’s incentives become misaligned...Instead, what you want to do is keep the overall goals aligned, and parallelize projects that point toward this goal by sharing reusable guides and processes. Especially with the advance of AI, it’s possible to minimize coordination costs by using AI for “rubber duck debugging” your ideas instead of relying on perfect alignment and consensus.
Protecting Privacy in the Age of AI: What Leaders Need to Know
A recent white paper titled ‘Rethinking Privacy in the AI Era’ from the Stanford Institute for Human-Centered Artificial Intelligence (HAI) highlights the growing data privacy concerns in the age of AI.
AI’s growing reliance on vast amounts of data raises critical concerns about data privacy. Leaders must understand the unique privacy risks posed by AI and take proactive steps to address them.
AI systems can lead to personal data leakage through their outputs, as research has shown that AI models can memorize and expose individual data points from their training sets. Additionally, AI enables inferences about individuals to be drawn based on data from other people, meaning one's privacy can be impacted even if their own data was not used to train an AI system.
While existing privacy regulations provide some protections, they lack sufficient governance mechanisms for the entire AI data supply chain, from initial collection to use in training models. To close this gap, a shift from the current "opt-out" model of data collection to an "opt-in" approach is needed. This can be achieved through strong privacy defaults and automated communication of user preferences.
Moreover, transparency and accountability must be ensured throughout the lifecycle of AI training datasets. This includes documenting data provenance and user consent, as well as assessing privacy risks at each stage of data processing. Data intermediaries can play a key role by enabling users to control how their data is used and facilitating the collective exercise of data rights.
Leaders should recognize that some constraints on data collection may be necessary to uphold privacy, even if they slow the pace of AI development. Policymakers must take proactive steps now to ensure an AI-driven world does not mean the unraveling of data privacy.
The path forward requires responsible data governance practices and public investment in privacy-preserving data resources and infrastructure. By taking these actions, it is possible to build a future in which the benefits of AI can be realized while still respecting and protecting individual privacy.
In the AI era, data privacy must be a top priority. Leaders across all sectors have a critical role to play in shaping practices and policies that uphold privacy as a fundamental value. The time to act is now.
Director's Corner: Should Boards Add Specialized Directors?
Spencer Stuart's Director Pulse Survey, conducted in April 2024, sheds light on the growing trend of boards seeking directors with specialized skills. The survey reveals that over the past three years, boards have added directors with expertise in areas such as finance, digital technology, cybersecurity, and AI to enhance their oversight capabilities.
The majority of respondents reported that these specialized directors have positively impacted the board's ability to oversee their respective areas of focus. Directors with cybersecurity and digital technology experience were perceived to have the most significant impact, followed by those with financial and AI expertise. This trend highlights the increasing importance of having board members who can effectively navigate the complex challenges facing today's corporations.
However, the survey also uncovered concerns about the potential drawbacks of relying too heavily on specialized directors. Many respondents emphasized the value of well-rounded board members who can contribute to discussions across multiple areas. Some expressed worries that an overemphasis on specialized skills could lead to a board of "one-trick ponies," detracting from the overall governance responsibilities of the board.
Respondents also suggested alternative ways for boards to access specialized expertise, such as advisory councils, external consultants, technology committees, and tapping into the knowledge of internal management teams. Ongoing director education was also highlighted as a valuable tool for helping boards stay informed about emerging trends and issues.
My View: Go For Generalist with Spikes
As a proponent of effective corporate governance, I believe that striking the right balance between generalist and specialist skills on the board is crucial. All board members should have a strong foundation in fiduciary duties, financial stewardship, and strategic oversight.
At the same time, I recognize the value of directors cultivating "spikes" of specialized knowledge in domains relevant to the business. These insights bring diverse perspectives to the boardroom and allow for deeper, more targeted questions that can help the board navigate complex challenges. Here, I emphasize that these directors are not mere 'specialists'; they are well-rounded directors even when taken away their 'specialties'. Their specialized knowledge should complement, not overshadow, their broader responsibilities and contributions to the board. And these directors won’t be happy if they are boxed into a specialty only.
Ultimately, the key to successful board composition lies in finding the right mix of skills, experiences, and perspectives to provide effective oversight and strategic guidance. By embracing both generalist and specialist skills, boards can position themselves to thrive in an increasingly complex and challenging business environment.
AI is Helping to Put Healthy Seafood on Your Table
Do you know that half of the world’s seafood now comes from aquaculture?
And do you know AI is already helping?
In Indonesia, the aquaculture sector is experiencing a technological revolution with the introduction of eFishery's AI copilot. This innovative tool is transforming the way local shrimp and fish farmers, such as Andriyono and Ira Nasihatul Husna, manage their operations. By providing real-time data on water quality and market prices, the AI copilot helps ensure optimal conditions for aquatic life, leading to healthier fish and shrimp. This not only boosts the farmers' productivity but also supports sustainable practices in the industry.
One More Thing
I am heading to Minneapolis, Minnesota to speak at a joint event by the National Association of Corporate Directors and Women Corporate Directors next week. Very excited about the trip…but there is a chance that the newsletter next week might be delayed.
Have a great week.
Joyce Li