As LLMs have improved, there is a lot of hype as to their impact on software development. Over the year IDE integrated tools like Github Copilot, Cursor have improved a lot as the underlying models improved. In addition to that there are completely standalone coding agents like Claude Code, GPT Codex and Google Jules/Antigravity have emerged with more capabilities. I haven’t worked a lot with the standalone coding agents, but have used Github copilot extensively and have seen the improvment.
So here are the concrete things that LLMs are helpful in improving the dev lifecycle from my experience. I’ve mostly used Claude Sonnet 4.5 and Opus 4.5 in my evaluation

Faster understanding of legacy code

LLMs can help us understand the workflows and technical organization of legacy code which speeds up understanding of the code. Initially they were poor in understanding large code bases because of context limitations. But once the agents learnt patterns to search, they execute multiple searches and calls and understand the code and can document the architecture for us to understand the flow. This helps us get up to speed with old code faster and also to become productive with new features or fixes in the code.

Peer to think through

LLMs take the rubber duck debugging concept much farther. You have much capable duck to talk through your issue, get concrete suggestions and alternative points to think through to debug/plan/re-architect your project.

Faster code construction

This one is obvious and the first usecase of the copilots. Now in the ‘agent’ mode given a requirement then can plan and one-shot the code required for implementing a feature. They can also compile the code and run them to check if the code is running. They can also execute tests to see if if the requirements are satisfied.

Tests

This brings us to the next big advantage. We can iteratively come up with the testcases that need to be created and the agent can write the tests and run them instantly increasing the confidence in the code not just that time but even for the future.

Faster Migration

Another avenue where LLMs shine is in the case where you have to migrate your code from one version to another like java 11 to java 17 or spring boot 2.x to 3.x or AWS SDK 1.x to 2.x etc, then these agents can help you migrate the code super fast. Just ask them to create a plan and then keep working through the plan as required. I needed to open some of the files and ask them to be converted and it did a great job and I was able to migrate a huge application in 1 day. This is a big win.

Error analysis

This is a corollary to the duck debugging, if you give an exception stack trace, the agent can identify the cause based on the exception and suggest some possible reasons and mitigations for the same.

Custom operational tools

With the faster speed of development, it is now possible to develop custom tools to help with manual operational stuff without having to spend a lot of time. It is my 20% project to develop new tools to improve our operational posture.

Pitfalls

Is everything rosy then? - Here are some of the things that don’t work as well or might be counterproductive

• It might have subtle bugs that are not immediately evident, so you have to review the code carefully and test all scenarios before merging the changes.

• You might forget syntax and rely too much on the LLM. Not an easy solution here but it might be ok?

• It can sometimes just comment out the code causing errors or failing tests after trying to fix them a few times, so it is important to keep looking at what the agents are doing.

• It might not stick to the existing architecture and want to make it similar to generic architecture it is trained on aka complex solutions. So it is imporant to make sure it complies with the architecture of the existing app.

• If it codes out a whole new app, then you might not really understand what all it is doing. You can ask it to document the architecture with diagrams to understand it.

Even with all these pitfalls, there is no doubt that it has double the productivity of software engineers if not more. It does make it more fun to work with code.

Having good unit/integration tests and blue/green deployments will help the teams take advantage of it. Marin Kleppman predicts that formal verification will go mainstream with AI programming. I think that is a great outcome though there is a learning curve.

Governments raises taxes in various ways

• income

• property taxes

• capital gains taxes

• sales/consumption taxes

• payroll taxes

• estate taxes

• corporate taxes

From a USA perspective, over the years there have been arguments about how one tax or the other incentivizes one thing or the other. For e.g. proponents of low capital gains taxes argue that it increases investments. And various governments have set various taxes at different levels. These different exemptions and levels have some unfair consequences - like long term capital gains tax at 15% is lower than 22% on income if someone makes more than 48k after exemptions. Various governments have tried to correct this inequality but it hasn’t really changed. And various governments have tried to reduce the taxes as they believe government is not really spending the tax money efficiently so they shouldn’t tax much. If governments don’t raise sufficiently money, then they will need to cut back on services or borrow money to pay for them. It seems taxing consumption with exemptions for food/medicine seems like a good way to tax since it incentivizes earning for consumption and it needs to be paid more by those who consume more.

Other countries tax more and provide more safety net to their citizens. There are some arguments that it decreases their economic dynamism. Tyler Cowen strongly argues that higher growth rate automatically lifts everyone though the wage growth relative to inflation in the past few years doesn’t prove that out.

What are the rational course of actions ?

• Governments should have enough money to at least provide the basic common services - police, fire, roads, parks, libraries and schools. So support this ability instead of cutting taxes to the bone.

• As an individual, asset owners benefit more, it is important to work hard to earn assets rather than relying on wages as soon as possible.

Sources: various economists(more from Tyler Cowen) /llms/youtube.

The Beginning

I’m sure you’ve used or at least heard about the Calm app for meditation. I had tried out various meditation apps like Headspace before Calm. When Calm came out I tried it out and it didn’t seem anything special. The Daily Calm was good but nothing that made me come back again and again.

The Breakthrough

Then I discovered sleep stories. They were an interesting novelty in the beginning but the stories were so boring that I didn’t give a second look.

My wife started liking the Daily calm before sleep and we started playing them daily before bed. As I started hearing them I got used to them and it felt like a good mental break.

Then one day I heard the story ‘Nordland Night Train’. It was very well produced and the narrator Erik Braa’s voice along with the gentle rattling sound of the train was a great wind down for sleep. It actually took me a few tries before I could get to the end of the story. Slowly I discovered a few more stories - Darjeeling Himalayan Railway by Bhavnisha Parmar, The Magic of Yurts by Candace Reardon - all of these stories started to grow on me. Slowly I started returning back to the app more and more.

The evolution

In the meantime they started adding more and more content. Started liking the daily trip with Jeff Warren. The Wisdom section also started having great gems. Now when I get up from sleep in the middle of the night, it is a reliable way to get me back to sleep.

The takeaway

I really liked the evolution of the app - how they started with minimum stuff and kept adding things that now the subscription seems valuable. It is a great lesson that any product developer can take. It is the slow way to success and if you care enough about a product, you can slowly iterate your way into adding enough value into it to make it successful.

And if you haven’t tried it already, I highly recommend trying it to both meditate and for helping you fall asleep. If you don’t meditate, learn why it is one of the best ways to improve your focus and attention.

I recently listened to a podcast from Mel Robbins with Dr. Amishi Jha a cognitive neuroscientist. She spoke about three types of attention with very interesting analogies.

The three types of attention

Flashlight - This is the first type of attention where we focus on on thing. When we are working on something with full focus or studying for an exam, we are using this kind of attention. We are using our attention ‘flashlight’ to focus on a particular thing. When we are focusing on that particular thing, attention on everything else is reduced.

Floodlight - This is the alerting attention, where we are alert and open to receive any inputs. The example she gave is when we are driving in a school zone. At that time, we are fully alert scanning large area at a time for child that could appear anytime. Here our attention ‘floodlight’ is spread across a large area and is open to receive any alerts. We are not really focusing on any particular place in general but are scanning everywhere.

Juggler - This it the third type of attention whose purpose is to oversee what we are doing like an executive. It is the thing that helps us juggle multiple things and keep them in the air. It helps with keeping us on track keeping our goals and behavior in sync. We use our juggler to override automatic tendencies and to update and revise a goal as new information comes in. This is also called the executive function.

How to strengthen them

She spoke about how we can strengthen them and the answer she gave was - meditation. She explained how meditation helps us strengthen all three types of attention at once.

When you focus on your breath, you’re using your flashlight. When you’re breathing you’re open to anything that arises, you notice it and let it go strengthening the floodlight part of the attention. Instead of letting a thought go, if you grab onto it and follow it, you realize and come back to the breath. This part is done by your juggler. Hence you’re strengthening it as well. So with meditation you’re actually strengthening all types of attention.
She claims that meditating for 12 minutes daily we can really improve our focus. The analogies are fascinating and got me back into meditation again.
Listen more at https://www.melrobbins.com/episode/episode-337/

Infrastructure

One of the first thing that is substandard is infrastructure. The road that leads to my home in India has gone from bad to worse in the past 10 years. Sometimes funds were allotted to improve the road before elections. Some work was done but it was washed away by the next rains. This small thing points out so many issues. All of them in no particular order

• There is no accountability. The contractor who laid the roads is not held accountable. And we wonder why? Is it because the R&B engineer doesn’t care or were there not enough funds allocated or were some of them “LOST” in between.

• There is no accountability part 2 - So the road is now not built but who can a citizen question about the roads? The local politician only visits during the elections. And he has a lot more things to take care of than a road. What is missing is a local accountable person. In the US we have a mayor who is responsible for local amenities in the city. He lives in the city and can be booted out if things don’t improve.

• But Indian municipal corporations also have a mayor who is elected indirectly by the people. So what gives? here is problem in the next layer. In the US the city has the ability to collect property taxes, raise bonds to pay for capital infrastructure improvements. In India taxes are collected once and devolved from center to states and states to local bodies which makes it difficult to be responsive to the needs of the people. So this centralization of power is a big issue.

Legal System

Everyday we hear news about government employees caught taking bribes. We hear news of land encroachments, about companies not following through on their promises. How many are resolved quickly. Most of the time nothing happens.
What recourse do we have in this case? Courts of course. But in India swift judgements don’t happen. Cases languish over years if not generations. I don’t have a first hand experience but based on my reading, it seems we don’t have enough judges for the amount of cases our system handles. This is another way accountability is avoided. US is a highly litigious society. Say what you may about the impacts, but it causes everyone to be extra careful. Yes there are problems because of this but I think having a swift and efficient legal system is important to resolve issues if they arise.

I don’t want to say these are happening purposefully and I want to acknowledge that we we are still a poor country and development is happening. But I think the lack of accountability is a structural issue and it needs to be solved in parallel to getting more resources. Otherwise, even if we get more resources, we will not see the quality of life increasing.
Yes there are many other issues, but having power de-centralized and a swift and efficient legal system can cause dramatic improvements. I hope we elect leaders who see that and take the required actions.

https://x.com/Devon_Eriksen_/status/1823878805812806030
https://x.com/melissajcave/status/1823887797842788783

It felt right but didn’t feel optimistic honestly. It also made me think if civilization means anything or is it still might makes right.

I had many conversations with GPTs and came away with the following points

• Yes guns act as a deterrent against using violence but that in a civilized society that shouldn’t be necessary. We can see that widespread availability of guns makes accidental crime happen a lot more. Hence true civilization should progress away from the use of force and rely more on trust in institutions.

• But what if institutions are corrupt and people are not rational. Even in that case though it might initially be helpful, soon it will devolve into mafias and shootouts. In this case again the most armed becomes the most powerful but now unaccountable and armed. The solution should be to improve the democratic institutions rather than being armed.

• There was an argument that firearms help overcome government tyranny. Though it seems true on the surface the power asymmetry between the government and citizens is too large. So if people revolt against government, it will result in too many casualties. It will result in prolonged violence. So in today’s world democracy, transparency and accountability are better safeguards.

• Next I investigated if civilized society can coexist with personal firearms. Switzerland and Finland have relatively higher levels of personal firearms but have low gun violence crime rate which shows it is compatible. But for that the institutions should be strong (better training/safeguards for gun ownership) and lower inequality through stronger social safety nets, strong culture of responsibility.

  This definitely makes me more optimistic about the role of institutions to make the use of force as minimal as possible. Yes sometimes use of force is necessary but it is better to focus on strong institutions, transparency and accountability.

Introduction

Have you ever lied down watching the stars and wondered what is there beyond the stars, beyond what our eyes could see? If you’re like me you’d have thought that universe extends towards infinity. It is never ending.

Infintity is confusing

But yesterday I watched a show on Netflix called - A trip to infinity which ended up blowing my assumptions.
They started by introducing the mathematical concept of infinity- where you an always add 1 and get a bigger number. There are some concepts that don’t fit neatly with the arithmetic rules.

For e.g.

And if you subtract two infinities you get 0

If you subtract infinity on both sides of equation 1 you get a unintuitive result

What does that mean?

Larger Infinity

What that result shows is that not all infinite sets are the same. You can compare infinite sets by showing one to one corresponce.
For e.g. Consider a set like 0.1, 0.2 etc. You can show a correspondence of

0.1 - 1

0.2 - 2

0.3 - 3 etc.. to show that they correspond 1:1 and hence these sets are equal

But there are other numbers like

are different kind of infinity. If you keep writing out the numbers of pi, you can keep creating an infinite set. This set cannot be mapped to the counting infinity set of 1,2,3, etc. Hence this is a different infinte set from the counting infinity set and is considered larger because it is not countable.

Apple in a box

This was the most interesting mind-blowing concept from the documentary. It goes like this - If you have an apple in a box from which nothing can escape and after a long time the apple gets rotten, in even more time, it turns into a gas and because of the chemical energy, the gas gets heated, then there will be a nuclear reaction forming plasma and subatomic particles and then slowly form matter. An apple has approximated 10^23 atoms. So they argue that there are only a finite states of matter and given enough time matter will cycle through all states and we might get to the original apple again. What that means is that everything in the world will form again given enough time.
The box is the universe and we are the apples. In that case, we will exist again. And if there are multiple universes there might be other people like us now. We may never meet them because those universes are far way and we cannot go faster than the speed of light.

This is called the poincare recurrence. This is called a ergodic hypothesis which assumes every microstate of the system is assumed overtime. There is interesting discussion about it here. The gist is that not every state is equally probable. But given infinite time, there is a chance that it would come back to the original state.

Is Universe finite or infinite?

We don’t really know if the universe is finite or infinite. As per Einstien’s theory of general relativity, universe is finite. It is open but not inifinite. It is like our earth where we keep going in one direction and we will reach that point again coming from another direction.

We never know because the observable universe in which light can reach in our observable time is limited. So we don’t really know whether it is finite or infinite.

We have observed that universe is expanding. The consequence of this is that we will be far from other stars in the galaxy and once our sun is all burnt out, all life on the planet will end.

This is a mind-blowing show to expand your mind. Definitely recommend this.

Feelings come from our thinking

In this book Joseph Nguyen writes that thinking is the root cause of suffering. Reality is the event that happened with no meaning or thinking or interpretation of it. Our feelings do not come from events but our thinking about them. Therefore we can only ever feel what we are thinking. This means we can change our feelings by changing our thinking.

Our natural state is joy, love, ecstasy, freedom and gratitude like a baby. Only the thinking we do will take us away from this state and make us anxious. On the other hand when we are in a state of flow, we don’t think much at all. So the more thinking we have going on, the more anxious we become.

Thoughts vs Thinking

He makes a distinction between thoughts and thinking. Thought is a noun, it arises naturally. Thinking is a verb and we do it. We can allow our thoughts to come and go through us while we minimize the thinking about the thoughts that emerge. The way to stop the thinking is not by trying to stop but by just becoming aware that we are thinking. Then the thoughts dissolve by themselves.

One of the analogies he gives is of murky water. When we let the water settle, the dirt will settle to the bottom and fresh clear water becomes separated. In the same way if we let our mind settle it without thinking will become clear.

Goals of Inspiration vs Goals of Desperation

No goal is good or bad it depends on where the goal is coming from. If the goal is coming from a place of desperation we feel a sense of scarcity and urgency. It feels like a burden. Even if we end up achieving it, we feel more empty than we did before. The way out of it is to create goals of inspiration. When we create goals of inspiration, we feel moved to create. We feel inspired and expansive and feels more like a calling than an obligation. It is like a powerful force of life coming from us that wants to be expressed through us to be manifesting in the physical world. That is why painters paint, dancers dance, writers write and singers sing even if they are not paid to do it.

One way to think about inspiring goals is to think about this question -

“If I had infinite money, already traveled the world, had no fear, and didn’t receive any recognition for what I do, what would I do or what would I create?”

Creating space for miracles

When you let go of thinking and start working on things that inspire you, you will be inspired. You will start trusting your intuition. When you’re faced with a challenge, create space by surrendering to the divine wisdom and trust that you will get an answer.

Once you start believing that universe is working for you not against you, miracles will happen.

This book really resonated with me, especially about having inspired goals. I strongly recommend this book to everyone.

• Question every requirement. All requirements should come with the name of the prson who made it.

• Delete delete and add back 10% if required

• automate only after deleting and simplifying the process. Don’t optimize a process that shouldn’t even exist.

• All technical managers should have hands-on experience

• comradery is dangerous

• meet with skip level managers not just the managers to know

• Look for people with right attitude. Skills can be taught attitude cannot be.

• Maniacal sense of urgency is the operating principle - Make progress - hardcore

• Only rules are the laws of physics. Everything else is a recommendation.

This project was developed during the LLM Agents Learning hackathon (https://rdi.berkeley.edu/llm-agents-hackathon/ ), where we set out to solve a critical challenge in the startup ecosystem using AI agents. The hackathon not only provided the platform but also offered valuable OpenAI credits that enabled us to experiment with different approaches.

The Problem Space

The tech startup ecosystem presents a unique hiring challenge. Startups, with their fast-paced nature, require specialized skills but often lack the visibility to attract top talent. Meanwhile, recent tech layoffs have created a pool of skilled candidates seeking new opportunities. Traditional hiring methods often fall short, creating a gap between available talent and startup needs.

This is where we saw an opportunity to leverage LLM agents to create a more intelligent matching system.

Our Journey: From Complex to Simple

Initial Approach: Experimenting with LLM Agents

Our journey began with ambitious plans to create a fully automated system using LLM agents. We explored three main approaches:

1. The Autogen Experiment

My first stop was Microsoft's Autogen framework. The promise was enticing – multiple AI agents working together to solve complex problems. I spent hours setting up agents to scrape startup information from various sources. The initial results were promising; watching Autogen generate scraping code autonomously was almost magical. But then reality hit: the websites we needed to scrape had robust bot protection. Autogen could write perfect code, but it couldn't overcome these real-world barriers.

2. Exploring Magentic-One

Undeterred, I pivoted to Magentic-One, another Microsoft framework that seemed perfect for our needs. The framework impressed me with its methodical approach – it would break down tasks, plan solutions, and generate code systematically. But as we tried to execute the code, we ran into a new problem. The framework got caught in frustrating loops, trying to resolve errors but never quite succeeding.

3. Claude's Computer Use

I experimented with Claude's computer use capabilities. It was fascinating to watch Claude navigate websites, access press releases, and convert screenshots to text. But this approach brought its own challenges – the extracted text was messy, filled with navigation elements and irrelevant content. Plus, the computational costs were adding up quickly.

The Hackathon Reality Check

Being in a hackathon setting significantly influenced our technical decisions. While we could have spent more time solving issues with Autogen we had to prioritize delivering a working prototype within the hackathon timeframe.

Pivoting to Optimal Solution

We ultimately opted for a more straightforward approach:

1. Data Collection

We switched to a more straightforward approach: manually accessing URLs and using the browser's reader view to capture clean text. It wasn't automated, but it gave us something invaluable – clean, reliable data.

2. Vector Database Implementation

We used OpenAI's Ada-002 to create embeddings of our company data and stored them in ChromaDB.

3. Matching

At the core of our solution is a nuanced matching system that goes beyond simple keyword matching. When a candidate uploads their resume, we use an LLM to parse their experience and preferences, convert this information into embeddings, and search our startup database for potential matches. The evaluation considers multiple weighted factors: industry alignment (35%), technical skills (25%), experience level (25%), and growth stage fit (15%). For example, a fintech startup might score highly for candidates with relevant industry experience, while the system also considers whether their years of experience match the company's stage – early-stage startups might prefer those with 0-5 years of experience, while growth-stage companies might seek more seasoned professionals. This multi-faceted approach consistently produced matches that resonated with both candidates and startups.

4. Outreach Package Generation

The final piece of our system focuses on turning matches into meaningful connections through personalized outreach. When a candidate selects a company, our LLM analyzes their resume, the company profile, and our matching scores to generate a tailored outreach package. The system creates a cover letter that emphasizes the strongest alignment points – whether that's specific technical expertise, industry experience, or growth stage fit. Alongside the letter, it identifies relevant company leaders to connect with based on role alignment. By automating this typically time-consuming process while maintaining a high degree of personalization, we help candidates make that crucial first impression count.

System Architecture

App Screenshots

We used Cursor AI coding tool for rapid prototyping. It reduced the development time and facilitated work with unfamiliar frameworks.

Lessons Learned

1. Claude Computer user works well but can’t scale.

2. Autogen and MagenticOne works through conversation between agents which is verbose and gets stuck in loops and generally not mature enough.

3. RAG worked well out of the box.

4. Agentic AI is an exciting space but it needs more stability and optimization.

Looking Forward

While our solution isn't as fully automated as initially envisioned, it successfully bridges the talent-startup gap. As LLM agent frameworks mature, we see potential for increasing automation while maintaining reliability.

Acknowledgments

A special thank you to Berkley RDI & LLM Agents MOOC for organizing this incredible hackathon and providing the educational foundation through their course. Their support, particularly the OpenAI credits, was instrumental in allowing us to experiment with different approaches and bring this project to life.

The project demonstrates that effective AI solutions don't always need to use the most advanced approaches - sometimes, thoughtfully combining simpler components can create powerful results. Sometimes, the best hackathon strategy is knowing when to simplify your approach to deliver working value.

Photo by Greg Rakozy on Unsplash

The Digital Trap

I crave to browse youtube shorts and it doesn’t take more than few seconds for me to get scrolling. I whip out my phone, and unlock it by FaceID, click on youtube and shorts and the short starts playing already. The only real thing I needed to do was to click on ‘Youtube’ as the phone was already out.

I want to write a paper. I can't (I tell myself) to do it on the phone. I have to get to my office[1], open my laptop[2]. Then I need to think of a topic to write about[3]. I need collect the information related to the paper[4], review it[5] and that is when I can get started.

It is fascinating how you can get your social-media fix anywhere but not do real work.

Act Fast and Take Control: The Power of the 5 Second Rule

Mel Robbin’s says that you have to act on a goal as soon as you think about it or your brain will stop it if you don’t physically move within 5 seconds.

By this rule you can see how our apps are perfectly designed to suck us in.

So what are we supposed to do here? You have a craving to scroll on social media and things are so easy that you do it without thinking.

Mindfulness: The Key to Breaking Habits

The solution is mindfulness. When you feel a craving to do something that is unhelpful, something to distract from real work, an addiction, something which feels good only temporarily - then the only thing you have to do is to take advantage of the five second rule. You get curious about the craving without acting. The operative words here are - “WITHOUT ACTING”. If you don’t act and stay with the craving it will pass.

That is how you can counteract your destructive addictions by just staying with it.

Pema Chodron, the great buddhist monk says, addictions are like wounds. When we scratch the wound and give in to our addictions, we don’t allow the wound to heal.

When we scratch the wound and give into our addictions we do not allow the wound to heal.

Pema Chodron

Learning to Stay with Discomfort

Here is another quote from Pema about how you have to get comfortable staying with the feelings without acting on them.

Staying with uncomfortable feelings and cravings is difficult. Practicing meditation can help. You can try any guided meditation apps for that. I use the Calm app and love it.

Transforming Thought into Action

The most effective way of dealing with addictions is to use the fact that thoughts come and go. So when a thought arises and that is not something you want to do, you just have to stay with it till it passes. But if it is something that you want to do, then jump on it right way before it passes. No wonder bias for action is such a big predictor for success.

The path to freedom

The digital age presents both challenges and amazing opportunities. By understanding the power of mindfulness and making conscious choices to act, we can harness the benefits of technology while minimizing its negative effects. Let's strive to create a world where we are masters of our own attention, not slaves to our devices.

Before I continue further if you’re not familiar with metabolic health, I want to introduce it. If you’re already familiar you can skip this section. 

Metabolic Health

Human metabolism is the set of chemical processes that occur within our bodies to convert food into energy and support various bodily functions. It involves the breakdown of nutrients, such as carbohydrates, fats, and proteins, to produce energy and other essential molecules. When we talk about metabolic health, we are primarily concerned with how effectively our bodies utilize and regulate energy and nutrients. When metabolism is healthy, your body can flexibly switch between all these energy sources (glucose, fat and protein) to generate energy. When it is unhealthy, it will be dependent primarily on glucose. Poor metabolic health can lead to various metabolic disorders, including obesity, type 2 diabetes, metabolic syndrome, and cardiovascular diseases. 

So here are some of the things that blew my mind 

Importance of muscles for metabolic health

• Skeletal muscle, or the muscle on our skeleton (contrasted with muscle in internal organs like heart or lungs)  is essential for glucose clearance and is responsible for taking in over 80% of glucose we eat as food (1, 2, 3). This means if you’re trying to improve your metabolic health, one of the best things you can do is to work out to build muscle.  The book goes on to talk about the things that are important in strength training like grip strength, loading and unloading carefully, pulling and hip hinging (exercises where you move around your hips like squats)

Importance of moderate intensity exercise for metabolic health

• We use more fats than glucose to generate energy when we are doing moderate intensity exercise (technically called zone 2). Zone 2 exercise refers to physical activity that is performed at a moderate intensity, which is typically around 70% of a person's maximum heart rate. 

• But when we exercise at high intensity, our muscles switch to burning glucose exclusively to fuel it. This blew my mind. I used to do any exercise at full throttle, which meant I was only working out on glucose. 

• If you use an Apple watch, for aerobic workouts, you can see the zone you’re working out in. You can walk uphill, bike or swim for training in zone 2. 

Silent decline of strength 

• Over the next 30 or 40 years, your muscle strength will decline by about 8 to 17 percent per decade—accelerating as time goes on. Now you probably don’t think  about briskly walking up a flight of stairs, but in no time, you will have trouble walking on a steep incline. This realization of how fast and how silent the decline was one of the most surprising things for me. 

• Peter suggests we start thinking about the things you want to do till you’re 100. This helps create a concrete goal and the metrics for various components you need to train for. The components he suggests are - strength, aerobic efficiency (zone 2), peak aerobic capacity (described below) and stability (described below).  We’ve already seen the first two above. Let’s see the other two below. 

Importance of peak aerobic capacity 

• In addition to building muscle and doing zone 2 cardio exercises, you also need to work on your peak fitness. This is measured in something called Vo2 max or the maximum amount of oxygen your body can use. This practically determines what you can and cannot do. You need Vo2(max) of 33 to be able to briskly climb stairs. The average 40 year old only  has a Vo2(max) of just above 30. This means it will already become difficult to climb stairs comfortably at 40. 

• The worst part is that this declines precipitously as we age. The good news is that you can train to improve it. 

• So starting with a high value and maintaining it high can mean that we are functionally younger. To train this, Peter suggests you do this  workout two times a week

• Ride a bike or jog for four minutes at the greatest pace you can sustain for this amount of time. Then ride or jog four minutes on easy setting which should be enough for your heart rate to come back down to below the 100 bpm. Repeat this four to six times and cool down. (32 min - 48 min). Make sure you always go below 100 bpm before increasing the pace again and also give enough time to warm up and cool down before you do this.

Importance of  stability 

• Attia says strength is important, but stability training should come first (like, six months of it before ever picking up a weight). Why? Stability is the secret sauce that allows you to create the most force in the safest manner possible.

• One of the key goals of stability training is to regain mental control, conscious or not over key muscles and body parts. By having that control, you load your body correctly. You’re able to react quickly when you slip and save yourself from a fall. You can do all your exercises without hurting yourself in the process. 

• This starts with breathing and specific exercises to get awareness of various parts in your body. Peter has a visual explanation for the things you can do to increase stability here Outlive Videos - Peter Attia 

You’re not eating enough Protein 

• The current RDA of protein is 0.8 gm per kg of body weight. It is the minimum amount to meet your basic nutritional requirements. But that is not the optimal amount. The optimal amount is around 2.2gm per kg of body weight. If you eat more protein than required, it is just excreted as urea but does not increase your weight. So eat more protein to build muscle, feel more full and be healthy.  

To train all four components, Peter recommends the following training plan each week:

• Stability: one hour, split into 5 to 10 minute blocks done before your other workouts

• Strength: three 45 to 60 minute full body workouts targeting all major muscle groups

• Aerobic Efficiency: four 45 to 60 minute zone 2 cardio workouts

• Anaerobic Performance: one 30 minute VO2 max workout

If you’re interested in staying stronger for longer, I would definitely recommend this book. I’m curious to know if these were surprising to you too. What surprising things have you learnt about health and fitness, let me know in the comments. 

Thanks to my wife Gayatri and my friend Saandeep for reading an early draft of this and providing feedback.

I recently read this book as part of a reading challenge. The author wrote this book in a peculiar format. A side character from one chapter becomes the narrator in the next chapter and so on. By doing this, the author is able to use different narration techniques in the same book. It also helps us understand a situation from different points of view. She is a real wordsmith. She writes one chapter as a field guide to a spy and another through messages between the characters. 

Bix, the founder of a successful social media company, is looking for new ideas to expand his company. As part of this, he attends a meeting at a college by a professor, Miranda Cline. Bix has used Miranda's original ideas about trust to build his social media company. In this meeting, he learns of a technology to transfer the consciousness of animals to the cloud. Bix takes this idea and creates a new product call Own your Consciousness. With this product you can transfer all your memories to the cloud. You can also change some recent traumatic memories and re-implant them back to your brain. Uploading your memories to the collective, gets you proportional access to others memories. 

This is very enticing and most of the world starts uploading their memories. Most people use this for personal and societal needs - to catch who was the kid that beat me up? What happened to that guy I shared a beer with at the Cafe? Who killed my friend etc. Chris remarks that the problem is that it is so attractive that people will give it everything. And the collective is omniscient. It is a Candy House - giving away your memories has benefits, but it can lead to unintended consequences. In fairy tale Hansel & Gretel, the witch uses the Candy House to lure them into the house. She feeds Hansel in the hopes of eating them. This reminded me of the aphorism in silicon valley, if you are not paying for the product, you are the product. 

Egan doesn't explore all the pros and cons of Own Your Unconscious. She tells us how the lives of different characters and how they live in the world with this tech. Charlene uses it to see her dad smoking pot at a commune when she was waiting for him to come home. Another daughter Roxy tries to relive a London trip with her father. Only it is leaves her with hurt. Later she uploads her memories and dies, finally feeling the connection with her dad. Alfred's character mocks the overperformance in social media. He hates it and takes to screaming in public to get an authentic response from people. 

Not everyone is enthusiasitc about sharing all the memories with the collective. Chris leads the resistance to this collective by building an app called Mondrain. He helps people elude from the collective by using proxies. They use fiction writers work to mimic them using their memories from the collective. 

Lincoln is a counter who helps identify the proxies and cut them. Lulu is a citizen spy with electronics called weevils embedded in her. Even after stint as a spy has ended, she faces trauma. She worries that the government is spying on her husband using the weevil implanted in her. There is Ames Hollander who runs an off the books store to remove weevils from former citizen spies. This helps them relax and not worry anymore about govt. spying. 

Bix's youngest son, Gregory doesn't share his father's interest about Own your unconscious. His father has shown his uploaded memories to the family many times. Those memories start with a blank space where his dad doesn't have any ideas. That interests Gregory. 

Later when he grows up, he comes to know that his father willed a lot of money to the Chris's resistance non-profit. He feels that he didn't know his Dad enough. One snowy day, he realizes that the blank vision his father started with was not blank. It was the human collective with each person doing different things. So it needed him to write his book to makes sense of all that. 

This novel is a huge panomara with lots of characters. There is no overarching narrative to the novel, except exploration of the lives of people around this technology. That is how Egan makes her point - stressing the importance of fiction in making sense of the human stories. 

“Only Gregory Bouton’s machine—this one, fiction—lets us roam with absolute freedom through the human collective. But knowing everything is too much like knowing nothing; without a story, it’s all just information. ” 

And that's why you need a narrator to connect the information into a narrative. 

These are the stories I liked the most - 

• Lincoln's story where he uses maths to understand the elusive thing that can make M fall in love with him. It is a clever expression of a math obsessed kid who loves M but can't express it to her in a way she understands.

• The story of Lulu, Molly, Chris and Colin when they share a close moment together on a pier. The teenage friendship, love and awakening moments was touching. 

• The story of Miles Hollanders' fall and redemption. Miles predicted that no one likes being right the first time. He then goes on to follow the script by falling from grace and then rising to be a state senator. 

• The story written in messages about Lulu's interview plan with Jazz Attenborough. It was clever and shows projects take shape.  

In the second book in the Ibis series, Amitav Ghosh writes about the situation in China before the start of the opium war. Through the various characters, he writes about the English, American and Indian traders in Canton. He continues exploring the lives of the various people who cames from Ibis to China, Mauritius and other islands nearby.

This made me think about the opium war and the lessons we can learn from it.

Britian had trade deficits with China. To reduce the deficits, they started pushing opium. Chinese get addicted to opium and they lose valuable bullion.

Conscentious Chinese mandarins appeal to both the British merchants and the Queen. Their appeal is logical since British ban it in their country. So how can their morals justify selling it in China?

They also ban bringing opium to China. Then the traders started smuggling opium by selling it from the outer islands. These smugglers then bring it to China in fast crab boats. They bribed the local officials and still continued the business. A new official comes and threatens the chinese traders. They burn these fast crab boats and stop any opium from the outer islands. They quarantine all the foreign traders in Guangzhou till they turn in all their opium.

They don't harm any merchants but destroy the opium. This angers the companies who have lost their profits from the opium trade.

They force the British government to act and go to war against China.

Due to the powerful ships and military equipment, British win. They extract unequal treaties from China and addict the Chinese even further. All this so they can earn money without any regard to the damage done by Opium.

The lesson here is that if you're not powerful, people will walk over you. Even if countries talk a big game about values and morals, you cant trust them. They will not hesitate to use their power in an unfair situation.

This causes an arms race as each country wants to be in a more powerful position than the other.

It also tells about how Indian traders from Bombay also took part in the opium trade.

Before the opium war, Canton (present day Guangzhou) was the only port open for foreign trade.

The arguments from the British side for the war are also interesting. They argued that China was not allowing free trade. It is stopping their people from getting the things they wanted. The book also provides an interesting counter for that. It argued that the British used the free trade argument when it suited them but not always. As an example, the British set the rules for procurement of ships in a way that only British firms can meet. This caused Indian shipbuilders to lose their business.

Head over to read more at http://ananthsreflections.blogspot.com

So I wanted to share a few traps that await you when you become a PM and how you can avoid them. 

You can read the first one - Don't be a warm body.  This post is about the second one. 

2. Don't buy into the hype machine

Your exec team have an idea. It fits with the strengths of the company. If this idea is successful, then your company will be wildly successful. They start selling this idea even before vetting it. They make plans about it. They talk about it in the all hands meeting. They bring you in as the PM for the product. 

You buy in to the idea and focus on building the product and getting it out to the market as soon as possible. 

You speak to your engineering team and they tell you it will take a long time to build. Given the hype inside the company, you don't find it difficult to get resources for the build. 

You speak to prospective customers and who appear super excited. If you ask for concrete committments everyone asks you to get back once the product is ready. After a long build time the product is now ready. You go back to the prospective customers who have expressed interest before. Now they are a little bit cool. They say the idea is good but it is lower in their list of priorities. One customer who helped you spec out the product says that the way you've built it is incompatiable. They ask for many changes to the product before they can integrate it. Now you realize that none of the prospecitve customers want to use the product. 

You wonder how can a product which is the future of a company fail so bad? 

Here what happened is that you've failed to test if the idea you have is solving a real problem? 

You allowed a costly build to happen without knowing if the product is useful. 

Understand that it is the PMs job to cut the hype and identify the hits from the duds. 

How to avoid this trap? 

• It is the job of the exec team and sales team to be rosy about prospects of the products and product ideas. 

• You as the PM should test the ideas as cheaply as possible to make a product usable, feasible and viable. 

• When faced with such a situation, you should speak with your customers. Use the Mom Test to understand if the problem you are solving is real

• Get concrete committments from interesting customers to understand the viability of the product. 

• If you are not able to get committments, it could be because of two reasons. First reason being it is not a real problem. Second reason they need to see the product to see the value. 

• If it is the first then  communicate it and help the company pivot to some other product. 

• If it is the second then use the testing strategies like mockups, static demos etc to test it out further. 

Understand that it is the PMs job to cut the hype and identify the hits from the duds. 

Head over to read more at http://ananthsreflections.blogspot.com

So I wanted to share a few traps that await you when you become a PM and how you can avoid them. 

1. Don't be a warm body 

When you join a team as a PM, especially in the initial stages, you can't choose your product. Your seniors assigns a product for you. Once you come upto speed, you notice that the prodcut doesn't make sense. Or atleast it is not the best use of your time and company resources. 

How can this happen?  - 

A company gets started with a single revenue generating product. They achieve product-market fit with that product. They start selling the product and hire hoping the good times continue. Then they do well and sature the market for that product. Now the company wants to build new products to keep growing and sustain their valuations. For that they do some brainstorming, listen to customers, think of the market to come up with new products. 

They confirm it with a few customers and they respond well. They get a product built and deliver to this customer. They are able to sell this product to a few customers but they can't grow it anymore. This gets assigned to a new PM, you. 

You realize this is not a good product or not a product that can grow. Here I'm not talking about product trying to solve a real problem but doesn't solve it completely. Here I'm talking about products you know there is no market. i.e. you are not solving anybody's problem. 

Ultimately, it is your job as the  PM of the product to own it. The sooner you can raise the alarm about this, the faster the company can pivot.

But the director has this product in his portfolio and he needs someone to continue building it. You speak to him and tell what you think. You suggest some new products to grow that business. Those are risky ideas but if they succeed there is a chance to grow the business. They need bigger budget. All valid reasons for why we cannot work on those new ideas. But since he has a budget to spend,   it compels you to work on ideas which you know don't amount to much. 

What can you do when you find yourself in this trap? 

• Once you realize this is a dead product, you owe it to speak the truth to power. 

• Create a deck or a document detailing your analysis and recommendations. 

• Tell them what outcomes you can expect if you continue in the same path. 

• Generate alternative ideas which you can persue/test out to meet the same objectives.

• Speak to other internal teams to get their reading of the situation. 

• Include bytes from customers and internal folks about the situation. 

• Use formal and informal forums to talk about what you found. 

• Your goal is to get everyone to acknowledge the situation. Then drive towards making a change to the next promising idea. 

Lots of times, everyone is busy to notice this. Ultimately, it is your job as the  PM of the product to own it. The sooner you can raise the alarm about this, the faster the company can pivot. 

Next, I will speak about how to the avoid the second trap - buying into the hype machine

Head over to read more at http://ananthsreflections.blogspot.com

I recently read the book Sea of Poppies written by Amitav Ghosh. It is a fictional story written in the backdrop of the Opium wars with China in 1830s. 

This is the first book in the Ibis Trilogy. Ibis is an American ship used for slave trade. After American abolished African slave trade in 1808, a British trader buys the ship. He plans to use it to transport indentured labor to Mauritius. Mauritius' sugarcane cultivation was labor intensive. Abolition of slavery caused acute labor shortage. Hence the plantation owners decided to import labor from India. 

This journey on Ibis to Mauritius brings together a group of people each one showing a different side of British colonialism. 

There is a  mulatto American who is on the crew of Ibis as it sails from Baltimore to Calcutta to espace racism.  There are the lascars who help with bringing the Ibis to Calcutta and later to Mauritius.  There is a chaste king of Raskhali whom the British sentence for forgery and send to Mauritius. There is a high caste lady who married an opium addict and wants to die by Sati but a low caste man saves her.  She is on the run and gets into the ship for indentured labors. There is the daughter of French botanist whom a Muslim lady raises and is at home in Bengali customs. When her father dies, she gets into the ship to escape a marriage. These and many other characters all come together in Ibis on the trip to Mauritius. 

Through the stories of these people, Ghosh tells us about the opium trade with China. There was great demand for tea, porcelian and silks from China in Britain and Europe. But the British didn't have anything that China liked to trade with them. Because of this, the British had to pay in silver and gold for the trade. They started exporting opium to pay for these goods. The Emporer banned trade of opium as the opium was addicting the Chinese. Even after the ban, opium was being smuggled. To stop it, the chinese captured the ports and destoryed the opium. Opium was one of the biggest revenue generators for the empire. So Britian went to war with China to allow them to trade opium. The Chinese lost this war and in a later opium war had to give away HongKong for lease. 

To get all this opium to sell, the British leaned on Indian farmers. They compelled the farmers to cultivate opium by entering into unfair advance contracts. This devasted the farmers as the opium cultivation was uprofitable. The land was being diverted from food crops so they couldn't grow their food anymore. This caused lot of farmers and poor to migrate to Mauritius as slave labor. 

It illustrated a part of colonial history I was not familiar with and wanted me to read more. The consequences of all these events are still visible in the world. India is a biggest legal opium producer for pharmaceutical purposes. Mauritius still has a large Indian origin population. China still uses the opium war as insult to national pride. 

Ghosh using different languages and dialects was able to make the story real. The book was an easy read and I couldn't put it down. If you have any interest in colonial history, it is a great read. 

Head over to read more at http://ananthsreflections.blogspot.com

• ChatGPT is an AI application that can answer any question you ask it. OpenAI trained it on human generated content up until 2021. It is something called a Large Language Model (LLM) where it generates some text based on the question.

• One surprising thing about chatGPT is how good it is at providing answers to questions. Normally, you would need to search Google and read through multiple pages to find an answer. ChatGPT provides you a single answer, as if it were an expert.  So you don't have to navigate through multiple pages to find the information you need.

• I've tested ChatGPT  by chatting with it about facts, code, soft skills, and creating plans. It gave great to good answers for most. 

• It is not perfect.  If you know what you're asking about you will know its limitations. Even with these limitations, it is immediately useful for lots of things. For me it is most useful for organizing things I already know, remember things that I knew of but forgot and helping with coding syntax instead of having to look up documentation. 

Where is the immediate impact

• It is a great companion to our work. It can be a great tutor when you're learning something.

• It writes code and creates essays so well, it seems that it will affect any test based on textual input. Think school essays, online coding tests etc.

• ChatGPT together with other AI tools like DALL-E, MidJourney will change creative fields. 

• You can know generate poems, stories, create artwork all through AI. 

• You can create animations and videos with prompts and scripts. 

• You can create designs and copy for your product. 

• AI generated content will explode. It will difficult to know what is AI created vs human generated.

• It will democratize all these skills so people who are bad at these can already do an OK job using these AI tools. 

• Good creators, developers, designers, story tellers, editors, animators will always have demand. Because after all AI is doing for now is regurgitating what these masters have been doing.

Good creators, developers, designers, story tellers, editors, animators will always have demand. Because after all AI is doing for now is regurgitating what these masters have been doing.

What about the future

• AI has exceeded human capacity in some games like Chess and GO. It makes moves which a human thinker can't think of. 

• With more progress, AI will also be able to create great content. The downside is that AI will also create useless content. It will again need humans to review the output to choose the best ones. And may be combine them to create an even better one.

With more progress, AI will also be able to create great content. The downside is that AI will also create useless content. It will again need humans to review the output to choose the best ones. And may be combine them to create an even better one.

Final Point

In summary, with this AI revolution, we will add a few more tools in our toolset to create things faster and better. 

Head over to read more at http://ananthsreflections.blogspot.com

But there are somethings that can't be learnt from books. I got a chance to work alongside some great PMs whose effortless way to get things done blew me away. I decided to learn as much as I can from those colleagues. I'm super thankful to have had the chance to work with them and learn from them.

I wanted to list the things I learnt in no particular order.

• Be proactive in approaching people and building relationships. - Even when they were new, they took the time to understand who can help, setup a meeting with them and introduce themselves. 

• Keep people updated, give and take help - Since they had built the relationship if they had any question about something, they didn't hesitate to reach out and ask for help. In the same vein if there was something they could help with, they would take the time to help with anything that they could help with.  

• Do the hard work to form an opinion and then express the opinion to rapidly test it and solidify it. - They take the time to understand the problem they are trying to solve, think about it from various angles and form an opinion. Then they would express the opinions any chance they get so that they can clarify and improve it rapidly. 

• Stay organized - They take the time to keep everything organized so they don't need to hunt for things when they need share or work on something. 

• Externalize your knowledge in writing and visuals to bring everyone to the same page - They would find ways to write out their opinions. As you know writing is a way to clarify your thinking. Another bonus is if you can draw simple diagrams showing all the parts of the system, that would help improve your thinking and also help explain things to people and bring everyone to the same page. 

• Always look out for opportunities to evangelize the products we are building. Use it to motivate the team and brag about the impact. - They always kept the vision and impact of the products they are building in mind. So they would also bring it in to anywhere required to motivate and also subtly brag about their and their teams impact. 

• Gently prod people to set high expectations - When someone suggests an unmotivated solution or a leisurely timeline, they would gently remind them about their effectiveness and help them set a lofty goal. 

• Co-create things with your team - To effectively work with others, they would start a draft and then use that to facilitate the discussion, all the while updating it live as others give their opinions and suggestions. This will both make everyone feel valued and help capture all the notes without wasting time later. 

• Keep the sales and account management team close. It helps to know the pulse of your products and helps you sell it better. - They worked closely with the sales and account management team to identify the beta customers, understand which customers would like the product, what possible impediments might be there in adoption etc which helps to keep the product moving smoothly from inception to launch. 

Head over to read more at http://ananthsreflections.blogspot.com

• People would like to have $100 now than $110 a month later. This tendency to favor present-self at the expense of future-self is called temporal discounting

• We can use mental time travel to make sure we don't discount our future so much. 

• Think about how a decision you make now will feel in 10 mins, 10 months and 10 years later to trigger this time travel to see the impact of the decisions. 

• Move regret in front of a decisions helps you avoid that regret and also prepares you about how to handle it if it comes to pass. 

Ticker tape vs zoom lens

• If you see the ticker tape of any stock, you will see many highs and lows. But if you zoom out for the same stock you can see that all these daily ups and downs smooth out. 

1. Our happiness is also like that. If you stress our daily up and downs you'll not be able to get the right perspective. Happiness is a long term game. 

2. If you're blowing something out of proportion, your learning pod can ask questions about whether this thing matters in the future to snap you out of it.

Path dependence of happiness 

• Happiness is path-dependent. If you win some money and lose it all, you'll be bitter compared to when you start by losing some money and then making it back. 

• Time travel helps you see that both outcomes are identical and one shouldn't be better or worse than the other.

Ulysses Contracts

• You can use pre-commitments or Ulysses contracts to prevent you from irrational behavior due to too much present focus. 

• You can use either limit access to things that aren't good for you or you can increase access to things you want to do more. 

Decision swear jar 

• Based on your or language of the people around you, you can identify if anyone is not being truth seeking. 

• These are generally overconfidence, being all knowing, blowing things out of proportion, generalizing, shooting the message, leading the witness, being overgenerous in editing our story, moaning or complaining, lack of self-compassion etc. 

• Being aware of this, you can shift yourself to be more truth-seeking. 

Mapping the future 

While planning it is important to identify the full range of futures, the probability of each. By making our decision making process explicit, we will have a more realistic view of the world and we can also prepare our response for each kind of future so we are better prepared. 

Backcasting

• Thinking backward from a goal and think about the steps, events that need to happen to reach that goal. 

• It helps us identify low probability events and work to improve them so that they are more likely to happen 

• It also helps us adapt the plan to the unfolding future as we know what we are expecting to happen at each time. 

• It also helps to identify inflection points for reevaluating the decisions.  

• Company can plan pre-committments for anything that we identified during the plan. 

Premortems

• In a premortem we start by asking what will cause this decision to fail. 

• This will help us to identify all the risks to the decision

• It will give voice to any person who is doubtful of the decision but doesn't want to come out and say it. 

• Research also found that my imagining a negative future we are likely to work to avoid that. 

Hindsight Bias

• Backcasting and premortems are complementary as they help us see the full space of the future. 

• If we don't see the full future space, we are likely to think that the outcome was bound to happen. 

• It becomes impossible to realistically evaluate the decisions and probabilities after. 

Head over to read more at http://ananthsreflections.blogspot.com