Notes from “The Algorithmic Leader” by Mike Walsh

Work backward from the future: focus on future customers, not existing ones

  • By understanding algorithms and how they shape our interactions and experiences as human beings, you can gain insight into the kinds of data-driven platforms and products that are likely to succeed in the future.
  • Unlike the platforms of the previous digital revolution, today’s machines are starting to teach themselves, and in doing so, they are gaining levels of mastery that match or surpass human beings’ abilities in specific fields or activities. Humans, however, are still in the driver’s seat when it comes to imagining ways to use machine intelligence to create experiences, transform organizations, and reinvent the world.
  • Algorithmic leaders can prepare for the future by understanding who their future customers are and what they might want. Masayoshi Son, CEO of SoftBank, is an example of an algorithmic leader who starts with a strong vision of the future and works backward from there.
  • Your kids are the frontrunner generation of the algorithmic age. Having grown up surrounded by AI embedded in all their products and applications, they will have a radically different set of expectations and perspectives about the way that the world should work. Learn from them.
  • The greatest driver of business value in the future will not be the algorithms that you use to optimize your operations and infrastructure, but those that create compelling experiences for your customers and clients. Use the Wheel of Algorithmic Experience - intentions, interactions, and identity - to imagine how algorithms might respond to what you customers want, as well as how they behave and see themselves. 

Aim for 10X not 10%: design operating model for multipliers, not margins 

  • The algorithmic age is a “winner-takes-all” era: the leaders who think big and invest in scale will be in the best position for continued success. Don’t fall into the trap of letting your digital transformation becomes digital incrementalism.
  • Great ideas from the past can hold you back from exploring new opportunities. Design your organization to be agile enough to look beyond past successes to embrace new capabilities without being slowed by rigid structures, hierarchies, and workflows.
  • True algorithmic innovation demands more than just serious computation and financial investment; it requires interesting data. Algorithms are only as good as the data you train them on. Let finding and developing compelling data be a core driver of your strategic plans.
  • Large organizations have an advantage when it comes to data if they are willing to leverage it. To do so requires assembling the right teams and partners and systematically identifying the parts of your business with the greatest potential for algorithmic reinvention.
  • Algorithmic technologies like the blockchain and smart contracts challenge the traditional structure of firms and raise the question of whether parts of them will even exist in the future. The future of your company may be no company at all.

Think computationally: analyze problems from first principles, not by analogy

  • If you want to improve the way you think, start by understanding the difference between reasoning by analogy and reasoning from first principles. Reasoning by analogy leads you to compare like with like - a limiting prospect. Using first principles allows you to take a problem apart and look at it again from the perspective of its fundamental truths. 
  • Like reasoning from first principles, computational thinking is a structured approach to problem solving that allows you to leverage data and algorithms to be more effective. When you think in this way, AI and machine learning platforms can help you devise smarter ways of predicting outcomes and generating insights.
  • The story of how Christopher Shallue and Andres Vanderburg used machine learning to discover a new exoplanet illustrates not only the potential of computational astronomy and other new algorithmic approaches to research, but also the importance of team design when it comes to best leveraging AI for solving problems.
  • A key barrier to computational thinking in your organization is algorithm aversion, or human mistrust of the recommendations made by the AI system. Sometimes the best way to get people to trust algorithms is to involve them in the design and management of an algorithmic safety process.
  • In the future, the most effective computational thinkers will be those who can directly express their ideas and execute their strategies in domain-specific programming languages. 

Embrace uncertainty: seek to be less wrong with time, rather than always being right

  • Adopting a probabilistic mindset allows you to be better prepared for the uncertainties and complexities of the algorithmic age. Rather than trying to be always right, probabilistic thinkers instead try to be less wrong with time.
  • Your organization’s ability to rapidly assimilate new data and insights will determine how well it manages uncertainty. Without a smarter way of running meetings, you will compromise your ability to effectively share information, manage projects, and make decisions.
  • Conducting a decision audit will allow you to distinguish the decisions that really matter from the ones that can be automated or delegated. The fewer unimportant decisions that humans need to make, the more we can fully engage with the important ones.
  • Assembling an algorithmic brain trust or your own mastermind group is a good way to share and systemize the use of data and AI in your organization.
  • The real value of running experiments is not to find solutions but to uncover better questions. AI will not automate innovation; it will help leaders focus on the issues and ideas worthy of further exploration.

Make culture your operating system: humanize and complexify, rather than standardize and simplify

  • Technology may have changed the hardware of your business, but culture is your true operating system. Creating an effective culture requires identifying and nurturing the right set of principles, rather than controlling people through processes. 
  • Algorithmic leaders can leverage data and machine learning to create a more autonomous and decentralized environment for their teams to work in. It is better to be a gardener who provides a fertile environment for growth than a prison guard whose job is to ensure compliance.
  • Clever team design is a good way to accelerate culture change. Aldo Denti’s pod teams at Johnson & Johnson are an example of how team structures can support innovation, agile management, and rapid development when breakthrough growth is required. 
  • Where people work is as important as how they work. In the future we will combine data science and computational design with behavioral science and anthropology in order to algorithmically reinvent our workspaces. 
  • It is difficult to embark on a journey of cultural transformation if you can’t have a fact-based conversation about what needs to change. Look for ways to collect data on how you work, and use this as a basis for hacking your culture.

Don’t work, design work: guided by user empowerment, rather than mere regulatory compliance

  1. The real job of an algorithmic leader is not to work but to design work. The question you should be asking is not “Are we getting results?” but “Do we have the right approach?”
  2. Look for the scaled-up solution. The story of Ali Parsa and Babylon Health illustrates not only how AI can disrupt a traditional industry, but also its role in building a service with global scale.
  3. The importance of designing work is not limited to finding innovative ways of doing things; it includes identifying, preserving, and replicating talent patterns, or the implicit knowledge and expertise of your best people before they leave or retire. 
  4. Empowering your teams to be citizen developers ensures that the people closest to the work can have the biggest say in designing it. Employees should be able to design work, even if they lack formal programming skills.
  5. A great example of using algorithms to design work is building a digital twin, or a digital version of your product, part, or process. In doing so, you may discover not only opportunities for automation but also entirely new business models.

Automate and elevate: ask whether have the right approach, rather than whether are getting results

  • Rather than simply eliminating jobs, automation changes them. The important question to ask yourself is not “When will my job disappear?” but “What is the new job inside my old one?”
  • As jobs transform, new skills will be required. Algorithmic leaders need to invest in their own capabilities to stay ahead of the AI revolution, and to remain relevant and valuable.
  • As illustrated by the story of the Google Legal Operations team, one effective way to elevate people is to create a team to rethink teams. A good operations team not only exists to gain greater efficiency but also seeks to constantly reinvent the function itself.
  • Automation is not only an opportunity to elevate your teams; it is also an invitation to profoundly reimagine what you do. Challenge yourself to explore things that you can do now but couldn’t do before the age of algorithms.
  • As we start automating more of the repetitive parts of daily work, the most valuable use of your time will be managing exceptions and finding nonlinear solutions to complex problems. 

If the answer is X, ask Y: manage by principles, rather than processes

  • As algorithms become more pervasive and capable of influencing human lives, algorithmic leaders need to be ready to ask, “Why?” Simply following the law is not an adequate moral compass in an age when laws are unable to keep pace with disruptive change. If you want to avoid breaching the trust of your users and customers, you have to find a way to act in their best interest.
  • Algorithms are not impartial. They reflect our biases and viewpoints. The best way to avoid automating discrimination is to embrace diversity. Surround yourself with people who can help you understand the cultural context of your systems and data.
  • While we will continue to worry about your inability to fully understand how machines make decisions, algorithmic leaders will need to find workable compromises to ensure that they can get the benefit of machine learning, even when a system’s recommendations are not entirely explicable. 
  • Dumbing down AI platforms to the extent that we can actually understand them may undermine their effectiveness. It is often more important to know why a particular optimum or target was chosen than to be able to explain the reasoning behind an algorithmic decision.
  • As organizations become more like algorithmic machines, we risk losing the ability to comprehend their end-to-end systems. Don’t lose sight of the complexities of your own business and the magic that lies in the details of how things work.

When in doubt, ask a human: believe that should automate and elevate, rather than automate and decimate

  • While technology allows companies to standardize and simplify their offerings, the most successful organizations in the algorithmic age will embrace the complexity of human behavior and translate it into individualized, immersive experiences.
  • Algorithmic systems lack common sense, so avoiding dangerous errors, bias, or unacceptable choices typically requires input based on human judgment. In the future, the ability to reach a human who is empowered to override an algorithmic decision may be vital to our safety or livelihood.
  • AI should not be just a product of a privileged few, but a platform to provide services for the many. The story of Ramya Joseph and her company, Pefin, illustrates how AI can deliver tailored financial advice to everyone. Equity is not the only issue. Given the importance of scale to data-driven systems, algorithmic leaders need to rethink their assumptions about the customer segments that they can profitably serve.
  • To properly address human needs, we need to develop the discipline of human-centered design for machine learning. For AI to be useful, it has to solve problems for people in a practical and empathetic way.
  • It might be able to automate work, but AI can’t override the importance of human relationships. By freeing us from the repetitive tasks, AI can give us the choice to focus on the human interactions that really matter.

Solve for purpose, not just profit: transform for purpose, not just profit

  • Work is more than just meeting our material needs; it is connected to our sense of identity and purpose. As we use more algorithms in the workplace, there is a risk that people will lose their connection to the value and rationale behind their labor. 
  • In the future, we will be either working for or on the algorithm. Avoiding the algorithmic inequality trap will require more than just new forms of taxation or regulation; it will also need a sustained investment in training and education by both companies and countries. 
  • Algorithmic management has the potential to revive the evils of Taylorism and, unless handled carefully, global social unrest and industrial action. 
  • We will all be working for talent platforms in the future, whether as freelancers or as flexible employees in a global organization. Leaders have an obligation to design platforms that they themselves would be comfortable using.
  • The journey to transformation begins and ends with you. Most of us started out as analogue leaders. It will take a conscious application of persistence, patience, and will to fulfill what the future wants from us: “to become algorithmic leaders”.


要查看或添加评论,请登录

社区洞察

其他会员也浏览了