• Focus on what you can control

• Your actions, judgments, and effort are yours. Everything else is noise.

• Accept reality as it is

• Resistance creates suffering; acceptance creates strength.

• Character over comfort

• A good life is built on virtue, not convenience.

• Prepare for adversity (premeditatio malorum)

• Expect obstacles—so you’re steady when they arrive.

• Respond, don’t react

• Pause before emotion turns into action.

• Practice voluntary discomfort

• Hardship trains gratitude and resilience.

• Detach from outcomes

• Do your best; let results follow as they may.

• Time is your most valuable asset

• Spend it deliberately—once gone, it never returns.

• Ego is the enemy

• Pride blinds judgment; humility sharpens it.

• Serve the common good

• We thrive by contributing beyond ourselves.

• Live each day as a complete life
• Act as if today is enough—because it is.

Ten Lessons for a Fragile World — Key Ideas

• Ego – Move from trying to beat the market or prove yourself, to supporting your life’s purpose and flowing quietly like a river.
• Humility – Recognize that everyone is imperfect; seeing others and yourself more generously reduces judgment and conflict.
• Equanimity – Don’t let life’s ups and downs sweep you off balance; maintain inner peace regardless of external changes (in markets or life).
• Attention – Treat your attention like a precious asset; avoid “junk scrolling” and notice what truly matters.
• Freedom – True freedom is the space between stimulus and reaction — the power to choose your response.
• Questions – Be comfortable living with uncertainty and unresolved questions; trust that profound answers unfold over time.
• Character – Your values and how you act matter more than external events — remain just and calm in all conditions.
• Resilience – The ability to think, wait, and fast builds strength; patience and discipline make you difficult to break.
• Love – Vulnerability and genuine connection (not defensive security) lead to growth and deeper fulfillment.
• Simple – True peace comes not from accumulating more, but from needing less — embracing patience and contentment (citing Kabir).

Making 2026 Goals Quantifiable

This framework converts values and lifelong lessons into measurable, repeatable actions—without losing depth or meaning.

1️⃣ The 3-Layer Quantification Model

For each goal, define:

1. Behavior Metric → What I repeatedly do
2. Cadence Metric → How often I do it
3. Outcome Check → How I know it’s working

This avoids vanity metrics and focuses on process + reflection.

2️⃣ Quantified 2026 Goals (Lesson → Metrics)

1. Humble Excellence (Ego & Humility)

Behavior: Seek feedback and share credit

• Metric: Ask for feedback from 1 stakeholder per month
• Cadence: 12 feedback conversations/year
• Outcome Check: ≥ 80% of feedback mentions collaboration & clarity

2. Emotional Equanimity (Equanimity & Freedom)

Behavior: Pause before reacting

• Metric: Practice the 5-minute pause rule in high-stress situations
• Cadence: Weekly journaling (≥40 weeks/year)
• Outcome Check: Regret score ≤ 2/5 for major decisions

3. Attention Protection (Attention)

Behavior: Deep, undistracted work

• Metric: 90 minutes of deep work/day
• Cadence: ≥ 200 deep-work days/year
• Outcome Check: Completion of 3–4 high-impact projects

4. Character Over Outcomes (Character)

Behavior: Values-aligned decision-making

• Metric: Write a 1-sentence “why” for major decisions
• Cadence: ≥ 2 documented decisions/month
• Outcome Check: Zero decisions reversed due to ethical discomfort

5. Living with Questions (Questions)

Behavior: Long-term thinking & reflection

• Metric: Read 1 reflective book/quarter
• Cadence: 4 books/year
• Outcome Check: Maintain a living questions journal with ≥ 12 entries

6. Quiet Resilience (Resilience)

Behavior: Discipline in body & mind

• Metrics:
  • Exercise ≥150 days/year
  • Meditation 10 minutes/day, ≥250 days/year
• Cadence: Weekly tracking
• Outcome Check: Energy level ≥ 4/5 for ≥ 70% of the year

7. Meaningful Relationships (Love)

Behavior: Presence & connection

• Metrics:
  • 1 device-free family meal/day
  • 1 deep conversation/week
• Cadence: ~300 meals + ≥40 conversations/year
• Outcome Check: Relationship satisfaction ≥ 4/5 (quarterly self-check)

8. Simplification (Simple)

Behavior: Reduce unnecessary commitments

• Metrics:
  • Limit to 3 professional priorities/quarter
  • Decline ≥ 1 non-essential commitment/month
• Cadence: Quarterly review
• Outcome Check: Stress level ≤ 2/5 most weeks

9. Redefining Success (Freedom)

Behavior: Use an internal scorecard

• Metric: Weekly reflection answering:
  Did I act with clarity, calm, and integrity?
• Cadence: ≥ 40 reflections/year
• Outcome Check: ≥ 80% “Yes” responses

10. Purpose Over Urgency (Integrated Goal)

Behavior: Intentional planning and review

• Metrics:
  • 1 annual theme
  • 4 quarterly resets
  • 40+ weekly reviews
• Cadence: Structured yearly rhythm
• Outcome Check: No major goal driven purely by fear or comparison

3️⃣ One-Page Quantified Scorecard

4️⃣ Final Principle

If a goal can’t be tracked weekly, it will be forgotten.
Quantification is not about pressure—it’s about awareness.

Have you ever moved to a new city and felt completely excited… and completely lost at the same time?

I’ve been living in Chennai for the past three and a half years. I had tons of adventurous experiences in Chennai, which I call STARS

• Sightseeing
• Temples
• Aquatics (Swimming)
• Road-learning (Driving)
• Savouring local food

I visited several new beaches and temples with my family, and the most fascinating one was the Kapaleeshwarar Temple because of its stunning architecture and the sheer number of people who visit it.

I also visited the aquarium with my family, where the sharks looked huge and intimidating. The underwater world was absolutely stunning, and the slow-moving eel looked surreal. Swimming is a new love I discovered after learning it recently. Floating in the water feels incredibly relaxing, and I often unwind by going for a swim.

One of the most interesting adventures in Chennai is facing the December floods. The roads fill up with water, and going to the office feels more like a rowing competition. You’re shocked to take an unexpected ‘second bath’ the moment you step outside. You even end up making new friends in the traffic. And when the power goes out, the complete darkness brings back old memories.

I will never forget the New Year’s celebration at Marina Beach. Thousands of people welcomed the year with incredible excitement. It felt like a grand festival, with hundreds of phones lit up the night as everyone captured the moment. The local snacks here are incredibly tasty. I’ve always been a coffee lover, and discovering karupatti coffee has become a refreshing escape from my regular routine.

In my journey, how can I forget to mention the sportsmanship the city has. It has given me various opportunities to gearup and run on various platforms. I have been participating in 21 km run for past three years and each time seeing 1000s of people all over gives immense joy. Exploring new places—beaches, temples, and even surviving the floods—gave me experiences that were both challenging and fun. These moments have now become part of my personal list of ‘must-do’ things in life.

(George Chouliaras et al., Booking.com Data Science)
Read the full article

Overview

The article proposes a framework to help organizations evaluate, prioritize, and adopt ML best practices by linking them to measurable quality attributes. Since not all practices are equally useful in every context, the framework provides a systematic way to select practices based on impact and constraints.

Key Components

1. Quality Model for ML

• Defines “quality sub-characteristics” relevant for ML systems.
• Examples: accuracy, robustness, scalability, testability, understandability, discoverability, compliance.

2. Best Practices Inventory

• Collected from literature + internal surveys.
• Examples:
  • Versioning data, model, and config
  • Continuous monitoring of deployed models
  • Automating workflows
  • Removing redundant features
  • Testing feature code
  • Containerized environments
  • Documentation
  • Modular/reusable code

3. Mapping Practices to Quality

• Each practice is scored 0–4 against how much it contributes to each sub-characteristic.
• Ratings were provided by ML engineers and scientists.

4. Prioritization / Optimization

• Organizations can’t adopt all practices.
• Optimization (greedy or brute force) is used to pick subsets of practices that maximize quality coverage under resource constraints.

Findings & Insights

• Coverage statistics:
  • ~5 practices → ~40% coverage
  • ~10 practices → ~70% coverage
  • ~24 practices → ~96% coverage (diminishing returns after this)
• Top impactful practices include:
  • Versioning for data, model, configs/scripts
  • Continuous monitoring of model behaviour
  • Automating ML workflows
  • Removing redundant features
  • Testing input-feature code
  • Shadow deployment
  • Constant performance measurement
  • Documentation
  • Modular & reusable code
• Not all quality goals are equally covered: weaker areas include scalability, discoverability, standards compliance.
• “Optimal” depends on what your organization values most (e.g., robustness vs. compliance).

Recommendations

• Don’t adopt practices blindly — prioritize based on your quality goals.
• Map practices to quality attributes to reveal overlaps and gaps.
• Adopt only the most impactful practices until diminishing returns set in.
• Use structured optimization methods to decide where to start.
• Revisit your practices regularly to adapt to changing priorities.

1. Origins of Plaid & Founding Vision

• Founded in 2012 to modernize financial services for the internet/mobile world.\
• Initially built consumer tools (dashboards, spend tracking) → hard to monetize.\
• Pivoted to B2B: enabling fintech apps to connect with banks via APIs and data access.

2. Product Expansion & Innovation

• Expanded beyond bank-linking to services like:
  • Plaid Check (identity verification)\
  • Plaid Beacon (fraud detection)\
  • Plaid Layer (permissions & data controls)\
• Focus on income/employment verification, analytics, fraud prevention, and alternative credit scoring.

3. Business & Strategic Moves

• Visa planned to acquire Plaid for ~$5B (2020) → deal blocked on antitrust grounds.\
• Post-deal, raised funding at higher valuations; later adjusted with market shifts.\
• Built strong leadership team (e.g., Jen Taylor) and emphasized engineering + fintech culture.

4. Challenges & Opportunities

• Enterprise sales in financial services: slow, compliance-heavy, relationship-driven.\
• Trust is central: Plaid is the “plumbing” of fintech → must be reliable, secure, private.\
• Exploring Generative AI applications (fraud detection, underwriting, risk modeling).

5. Advice & Reflections

• Early customers: “free radicals” → influential users who evangelize products across startups.\
• Product mindset: start narrow, solve deeply, expand gradually. Ship early, iterate fast.\
• Leadership lessons: Founders must evolve — scaling teams, managing culture, keeping vision alive.

Sources:\

• a16z Podcast: My First 16 with Zach Perret\
• Acquired.fm: Plaid & Visa Deal\
• This Week in Fintech – Plaid Podcast Notes\
• Stratechery Interview with Zach Perret

What is this book about

• Part of the Great Mental Models series, designed to provide mental tools for clearer thinking and better decisions.
• Volume 4 focuses on Economics and Art.
• Includes over 24 models — from markets and incentives to genre, plot, and representation.

Key Mental Models (Examples)

Model What it is How it helps / Why it matters ———————– ———————————- ———————– Creative Innovation often destroys old Helps spot disruption Destruction structures, creating new ones. opportunities and anticipate losses.

Monopoly vs Market structures: monopoly powers Strategy in Competition vs competitive pressures. business/life: niche vs competition.

Representation How ideas are communicated Strong representation visually or symbolically. makes ideas persuasive and lasting.

Genre Socially constructed categories in Knowing the “rules” art or products. helps use or break them creatively.

Plot Narrative structure of Makes communication beginning–conflict–resolution. compelling and memorable.

Optimization vs Efficiency vs flexibility in Helps balance Adaptation changing environments. perfection with robustness.

Incentives Drivers of behavior, Essential for designing costs/benefits. policies, teams, products. ———————————————————————————-

Themes & Big Ideas

• Neutrality of models: Not good/bad; their use matters.\
• Economics + Art: Combining both domains enriches perspective.\
• Value creation: Representation, framing, and genre shape perceived value.\
• Trade-offs: Efficiency vs flexibility, competition vs monopoly, innovation vs stability.\
• Framing & narrative: How ideas are told matters as much as the content.

Applications

• Business decisions: Use incentives, competition, externalities.\
• Communication: Apply plot, representation, framing.\
• Innovation: Use creative destruction, bend genres, stay adaptable.\
• Personal choices: Weigh trade-offs, incentives, and context.

Critiques

• Risk of overusing models (“hammer and nail” problem).\
• Some models contradict; context matters.\
• Culture and audience differences change how models apply.

Summary in short

This book combines economics and art models to help you understand mechanics of markets and incentives, as well as perception, creativity, and communication. By practicing both, you can make better decisions, be more persuasive, and expand creativity.

🇮🇳 India’s Semiconductor Push

• Tata Group and Taiwan’s Powerchip are building India’s first commercial semiconductor fab in Gujarat with $11 billion investment.
• The factory is part of a broader plan to create a “Semiconductor City” and reduce dependence on Chinese chip imports.
• This follows India’s $10 billion incentive scheme launched in 2021 to promote domestic chipmaking, prompted by COVID-era chip shortages and security concerns.

🧩 Scope of Current Projects

• Five major chip-related projects have been approved so far.
• Most focus on assembly, testing, and packaging (ATP)—a labor-intensive but less technically complex segment.
• Tata’s fab is the only one attempting from-scratch chip production, though not at cutting-edge levels (e.g., not comparable to Taiwan’s TSMC).
• Chips produced will target automobiles, home appliances, and basic smartphones, including Tata vehicles.

💸 Subsidies and Skepticism

• Central government covers 50% of fab project costs; states add another 20–25%, raising concerns about cost-effectiveness and subsidy overload.
• Critics argue that there may be more efficient ways to create jobs than heavy investment in chip manufacturing.

🏭 Challenges and Opportunities

• India originally hoped to attract top-tier players like Intel, but has had to temper expectations.
• Experts like Rakesh Kumar and Pranay Kotasthane support India’s cautious, methodical approach to building chip capacity.
• There’s a risk of future protectionism, where companies may be forced to use subpar local chips.

🧠 Missed Opportunity in Chip Design

• India has a strong presence in chip design talent—about 125,000 Indian designers, mostly for foreign firms.
• Local chip design startups remain underdeveloped, despite government support.
• Investing in chip design might offer faster returns than building full-scale fabs.

🌍 Geopolitical Context

• Rising U.S.–China trade tensions could favor Indian exports over Chinese chips.
• However, potential global economic disruptions from Trump-era tariffs may dampen international investment enthusiasm.

🔍 Final Thought

• India’s semiconductor ambitions are ambitious but realistic in stages. Long-term success will depend on continued policy support, a balanced focus on design and manufacturing, and global economic cooperation.

• Leadership vs. Management:

Tao shares an early career realization that leadership is more than just being nice—it requires study. Simon emphasizes that leadership is not about rank but responsibility. People don’t want to be managed; they want to be led. Management involves handling processes, while leadership is about fostering growth in people.

• Leadership as a Skill:

Leadership, like any other discipline, must be learned and practiced. Many professionals are trained for their jobs but receive little leadership training when promoted. Great leaders continuously study leadership through books, discussions, and reflection.

• Psychological Safety & Trust:

People thrive in environments where they feel safe to admit mistakes, ask for help, and share concerns without fear of retribution. Being “nice” isn’t enough—constructive feedback is essential for growth. Leaders must lead by example, demonstrating vulnerability by admitting mistakes and seeking help when needed.

• Building Psychological Safety:

Trust isn’t instant; it develops over time through consistent actions. Leaders set the tone—if they are defensive, dismissive, or authoritarian, they discourage openness. Encouraging feedback and open discussion helps create a culture of trust and accountability.

Key Leadership Example – Peter Dimaggio:

Tao recalls a great leader, Peter, who never directly answered questions but instead pushed employees to think for themselves. This approach developed self-reliance and critical thinking while maintaining a supportive environment.

Extinguishing Psychological Safety:

Leaders destroy psychological safety by:
Ignoring or punishing vulnerability (e.g., dismissing concerns or ridiculing mistakes). Fostering fear-based leadership, where employees hesitate to speak up. Being inconsistent or unpredictable, leading to distrust.

Conclusion:
The session underscores that leadership is a responsibility, not a title, and must be actively cultivated. Psychological safety and trust are crucial to fostering innovation and engagement in teams. Leaders must set the example by embracing continuous learning, vulnerability, and open dialogue.

Below are few reasons:-

• Model Development:- We might take lot of time to build and then during the process business requirements change therefore, the need of the model is no more.
• Model Deployment:- Once we development, then comes the turn to deploy the model in the business processes. This requires special skills such building data pipelines, model hosting and scaling the platforms. It is possible that we don’t have right people or skills available in the organisation.
• Model Approval:- Once, it is deployed, then we need to check the performance of the models to meet the required business expecactions. If it doesnt meet the required business expectation then again it is possible that it wont be approved by Stakeholders.

We can navigate the challenges to create the business impact through predictive models. One of the challenging problem in creating the impact is to build the robust machine learning models which translates to required business expectation. This can be addressed from more theoretical approach as well. This paper talks about the Robust Machine Learning.