15 reasons Why reading books should be part of your Life

1. Knowledge Highway: Books offer a vast reservoir of knowledge on virtually any topic imaginable. Dive deep into history, science, philosophy, or explore new hobbies and interests.

2. Enhanced Vocabulary: Regular reading exposes you to a wider range of vocabulary, improving your communication skills and comprehension.

3. Memory Boost: Studies suggest that reading can help sharpen your memory and cognitive function, keeping your mind active and engaged.

4. Stress Reduction: Curling up with a good book can be a form of mental escape, offering a temporary reprieve from daily anxieties and a chance to unwind.

5. Improved Focus and Concentration: In today's fast-paced world filled with distractions, reading strengthens your ability to focus and concentrate for extended periods.

6. Empathy and Perspective: Stepping into the shoes of fictional characters allows you to develop empathy and gain a deeper understanding of different perspectives.

7. Enhanced Creativity: Reading exposes you to new ideas and thought processes, potentially sparking your own creativity and problem-solving skills.

8. Stronger Writing Skills: Immersing yourself in well-written prose can improve your writing style, sentence structure, and overall communication clarity.

9. Improved Sleep Quality: Swap screen time for a book before bed. The calming nature of reading can help you relax and unwind, promoting better sleep quality.

10. A Portal to New Worlds: Books transport you to different times, places, and realities. Experience thrilling adventures, historical events, or fantastical journeys from the comfort of your armchair.

11. Lifelong Learning: Reading is a journey of continuous learning and self-discovery. There's always something new to learn, regardless of your age or interests.

12. Conversation Starter: Books provide a treasure trove of topics for conversation, fostering connections and enriching discussions with others.

13. Family Bonding: Reading together as a family creates lasting memories, strengthens relationships, and instills a love for reading in younger generations.

14. Budget-Friendly Entertainment: Compared to other forms of entertainment, books offer a cost-effective way to enjoy leisure time and broaden your horizons.

15. A Gift that Keeps Giving: Books are timeless presents that can be revisited and cherished for years to come, offering endless enjoyment and intellectual stimulation.

Weight Management Consideration

You can Eat less, Exercise more and still miss the bigger picture.

Over the years, I have met thousands of people struggling with obesity. Some have tried every diet under the sun. Some practically live in the gym. Others eat so little that I wonder whether they've mistaken hunger for discipline. Yet, the weighing scale refuses to cooperate.

Because obesity is much more than calories in and calories out. It involves several interconnected dimensions.

1. Nutrition: What you eat, how much you eat and how your body responds to food.

2. Exercise: Building muscle, stamina, mobility and cardiovascular fitness. Not just burning calories.

3. Metabolic health: Insulin resistance, hormonal influences, sleep and other underlying factors.

4. Emotional health: Stress, emotional eating and your relationship with food.

5. Identity: Moving from someone constantly trying to lose weight to someone who actively lives a healthy, physically capable life.

There is another important nuance. Diet and exercise don't always deserve equal attention at every stage. Initially, dietary changes may need greater emphasis. As you progress, resistance training, muscle development and physical activity become increasingly important. Your body changes. Your requirements change. Your approach must evolve too. Which is why copying your neighbor's diet or your gym trainer's workout isn't necessarily a brilliant idea!

Think in terms of 6–12 months of consistent lifestyle changes, not crash diets and ambitious Monday morning resolutions. It's building a stronger, healthier and more capable body, along with the habits to sustain it. The goal isn't to spend your entire life trying to lose weight. It's to build a life where looking after your health becomes second nature.

Fruits & Diabetes

Are you afraid of eating fruits because you have Diabetes?

In practice, we see two very different kinds of people:

Some people with diabetes have completely stopped eating fruits. They believe every fruit will shoot up their blood sugar.

And then there are others who have become so enthusiastic about healthy eating that their entire breakfast consists of fruits!

Both approaches need a rethink. You don't have to give up fruits just because you have diabetes. You need to understand which fruits to eat, how much to eat and how to eat them. Winter gives us some wonderful options. Let me share four of my favorites.

1. Guava (Amrood)

This humble fruit is a powerhouse of soluble fiber. It's fiber that helps slow down glucose absorption in the intestine. Guava is also an excellent source of Vitamin C. One medium-sized guava can take care of a substantial portion of your daily Vitamin C requirement.

2. Amla (Indian Gooseberry)

If there is one winter fruit strongly recommended to include in your diet, it is amla. It is rich in Vitamin C and antioxidant compounds that support metabolic health. You can have one medium-sized amla with a little salt. If you prefer, 2–3 teaspoons of unsweetened amla juice are another option.

3. Ber (Indian Jujube)

Our good old ber! Remember the song 'Ber Lo, Ber Lo'? This humble fruit is often overlooked, but it contains fiber and beneficial antioxidant compounds. And don't dismiss the newer varieties. Apple ber, for instance, is wonderfully crunchy and delicious. Enjoy 5–6 small ber fruits, keeping your overall portion in mind.

4. Pear (Nashpati)

Pear is rich in water and pectin, a soluble fiber that helps slow down glucose absorption. Have a medium-sized ripe pear with its skin. Don't peel away the goodness!

Now, there is one more thing I want you to remember. Please don't convert these fruits into juices. Eat the whole fruit. The fiber is an important part of the benefit. And don't forget, the goal of diabetes management is not to make your food choices miserable. It is to make them smarter.

You can enjoy seasonal fruits while managing diabetes. Just be mindful of portions and monitor how your body responds.

Swelling & Diabetes

Persistent Swelling should never be dismissed as just another symptom of Diabetes. If you have Diabetes and notice swelling around your ankles or feet, there could be several reasons behind it. Some are relatively simple to manage, while others need medical attention.

Let me explain two common reasons:

1. Your medications could be responsible

Certain diabetes medicines like pioglitazone and sometimes insulin, BP medicines like amlodipine, certain painkillers, steroids, and even medications like pregabalin or gabapentin used for neuropathy can cause swelling in your feet.

The good news is that medication-related swelling can often improve once your doctor reviews and adjusts your treatment.

Please don’t stop or change any medication on your own.

2. Your Lymphatic drainage or Venous circulation may not be working efficiently

We all know that the heart pumps blood through arteries and veins carry it back.

But there is another important system called the lymphatic system, which helps remove excess fluid from our tissues.

Unlike blood circulation, lymphatic drainage depends heavily on muscle movement. This explains why some people notice swelling after sitting in a car for hours, taking a long flight, or spending most of the day without moving.

My advice is

• Get up and walk for 2–3 minutes every 30–60 minutes.
• Practice calf raises and ankle rotations.
• Elevate your legs when resting, if comfortable and medically appropriate.

If your veins are not functioning properly or you notice prominent varicose veins, please get them examined. And remember, ankle swelling in someone with diabetes can sometimes indicate kidney or heart problems, too. Sudden swelling, especially in one leg, or swelling accompanied by breathlessness or chest pain needs urgent medical attention. The earlier we identify the cause, the better we can address it.

Signs your daily workout routine tell?

Some changes never show up on a scale. If you’re a man trying to lose Fat & Build Muscle, I’d ask you 5 questions before suggesting a change to your workout. 

1. Are you getting stronger?

Your body may look the same even when you’re making progress. Can you lift a little more than you could a month ago? If that hasn’t changed either, I’d look at how you train and how much protein you eat across the day.

2. Are you tired even on rest days?

Feeling tired after a tough session is one thing. If you’re still exhausted two or three days later and it keeps happening, speak to your doctor. Low iron is one possible reason worth checking.

3. Does every meal feel incomplete without something sweet?

Wanting a piece of mithai after lunch doesn’t tell me you’re short on protein. I’d ask a simpler question. Are you still hungry, or do you just enjoy ending lunch that way?

4. Are you hungry again an hour after eating?

Think back to your last lunch. Was it enough for the day you were having? Alongside your roti or rice, was there a good serving of dal, curd, paneer, eggs or fish?

5. Is your weight dropping along with your strength?

Perhaps your sleeves feel looser, your waistband fits the same and your usual weights feel heavier. That’s when I’d review your meals and strength training to see how well you’re holding on to muscle.

These patterns are worth paying attention to. On their own, they can’t tell you whether you’re low in protein, vitamins or minerals.

Indian Seasonal Greens

Still Eating Only Palak & Methi every Winter? You are missing out on some wonderful seasonal greens!

Every winter, we bring home palak and methi almost out of habit. Both are nutritious, no doubt. But our Indian sabzi mandis have so much more to offer. And I often feel that some of our most nutritious traditional foods are the ones we tend to overlook. Here are 6 leafy greens worth adding to your winter meals.

1. Sarson ka Saag (Mustard Greens): Rich in vitamins A, C and K, along with fiber. Pair it with makki ki roti for that traditional winter favourite.

2. Gongura (Sorrel Leaves): Popular in Andhra kitchens, these tangy leaves provide iron, folate and antioxidants. Try them in chutneys or dals for a delicious change.

3. Bathua (Chenopodium Leaves): A wonderful seasonal green containing iron, calcium, folate and vitamins A and C. Steam it, add a little jeera and enjoy it with roti.

4. Chakwat (Wild Spinach): A good source of fiber and minerals, including calcium and iron. Cook it with simple spices for an easy, nutritious sabzi.

5. Moringa Leaves (Drumstick Leaves): Contain plant protein, calcium, iron and vitamins A and C. You can easily add them to dal, soups or everyday sabzis.

6. Amaranth Leaves (Chaulai): Provide vitamin A, folate, iron and calcium. Lightly cook them with garlic and simple spices for a wholesome meal.

Now, my suggestion is not to replace palak and methi. Instead, bring variety to your plate. Different leafy greens offer different combinations of vitamins, minerals, fiber and protective plant compounds.

Rotating them through the week is a simple way to improve the nutritional variety of your meals. Most of these have been part of traditional Indian cooking for generations. We just need to bring them back to our kitchens.

Magnesium & Diabetes

What if one missing Mineral is making your Diabetes harder to Manage? The mineral I’m talking about is magnesium. Your body uses it to help regulate blood sugar. A high sugar reading alone can’t tell you whether you’re low in magnesium, but it’s worth asking how often foods rich in it find their way onto your plate.

One place I’d start is the little dabba of saunf that comes out after dinner. I like that dabba for a simple reason. Nobody needs a reminder to open it. Someone passes it around, you take a spoonful, and the conversation carries on. The next time you fill it, try mixing pumpkin seeds, flax seeds (alsi) and sesame seeds (til) with your usual plain saunf. All three seeds contain magnesium. The pumpkin seeds give the mukhwas a little crunch, and you still get that familiar sweetness from the saunf. Start with a teaspoon of the mix after dinner.

There are other places magnesium can fit into your day, too. If you enjoy evening chai, have a few badam with it. If sprouted moong usal is already a favourite at home, that’s another source you can include more often. When I talk to someone about food and diabetes, I like to ask, “What does a usual day of eating look like in your home?” That answer tells me where a small change might actually work.

Sometimes it’s as simple as what goes into the saunf dabba the next time you refill it.

Google AI, ML, RAG and Agentic-AI Interview Questions (2026)

Google does not publish its interview questions. Every source behind this list is a prep vendor or an anonymous review site (Glassdoor, Blind, LeetCode Discuss). Use the topics as the signal and treat the exact wording as approximate.

How to read the tags

• [R] Reported: a candidate post or review, or a coaching vendor citing such reports.
• [V] Vendor-tagged: a prep site labels it as Google; the underlying report is not visible.
• [P] Prep-only: not Google-specific, or our own synthesis. Useful for breadth.
• [~] Practice: a practice scenario we could not trace to any source page.

The short version: the classic ML system-design questions are credibly reported by Google candidates. Most RAG, agent, and LLM-serving prompts online are good practice scenarios, not confirmed transcripts. Coding and a resume-driven ML deep dive carry as much weight as system design.

Verified ML system-design questions

• Build, train, and deploy a system that detects whether multimedia or advertising content violates the terms of service or is offensive. [R]
• Design autocomplete and/or spell-check for a mobile device. [R]
• Design autocomplete and/or automatic replies for email. [R]
• Design the YouTube recommendation system. [R]

Follow-up interviewers add

• How would you optimize prediction throughput for an RNN-based model? [R]
• Which loss function would you optimize, and why? [R]
• What data would you collect, and why? [R]
• How do you avoid bias and feedback loops? [R]
• How do you handle a corrupt model or an incorrect training batch? [R]

Further reported prompts

• Design a job-recommendation system. [R]
• Design type-ahead search. [R]
• Explain precision and recall. [R]
• Explain root mean square error. [R]

Additional Google-tagged ML questions

ML system design

• Design a content-moderation system. [R]
• Design a system that processes traffic data. [R]
• Detect ad boxes in YouTube videos, then code the frame ranges and positions. [R]
• Predict which links to show from user, timestamp, query and click history. [R]
• Design a real-time recommendation system for a consumer product. [V]
• Design a product or video recommender. [V]
• Design a feedback-driven recommender (contextual bandit, online learning). [V]
• Design large-scale near-duplicate video detection. [V]
• Design obstacle detection for a self-driving car. [V]
• Predict a reaction factor from molecule pairs (tabular, end-to-end). [V]
• Split a matrix multiplication across two GPUs. [V]
• Design an ML experiment tracking and analysis platform. [V]
• Design Gmail spam detection; Search autocomplete; a petabyte-scale training pipeline. [P]

ML depth and LLM fundamentals

• Walk through the full LLM lifecycle: data filtering, pretraining objectives, architecture, alignment. [V]
• Explain the Transformer and its variants in depth. [V]
• Logistic regression, Naive Bayes, multi-class metrics, bagging vs boosting. [V]
• Compare regularisation methods and when to use each. [V]
• Compare tokenization approaches; design LLM-based recommendation parts. [V]
• Explain GRPO-style reinforcement learning for diffusion models. [V]
• Implement a multi-head attention class, then discuss current architectures. [R]
• Sample efficiently from a multinomial distribution. [R]
• Implement k-means or logistic regression from scratch; convolution and batch norm in pure Python. [R]
• Precision and recall, over- and underfitting, regularisation, dropout. [R]
• Handling overfitting, when to stop training, and deploying on GCP or AWS. [R]
• Walk through a modeling challenge: detection, alternatives, fix, evidence. [V]
• Why might a model that is good offline fail online? [P]
• Your A/B test won, but your manager fears long-term harm. [V]

RAG questions

• Design a RAG question-answering bot: chunking, embeddings, vector-store choice, retrieval, response synthesis. [V]
• Design an enterprise RAG pipeline over millions of PDFs with retrieval under 50 ms and no context overflow. [V]
• Evaluate retrieval precision and recall separately from generation quality. [V]
• Compare dense and sparse retrieval; build and tune a hybrid system for low latency. [V]
• Detect, quantify and mitigate hallucination and prompt injection in a multi-step workflow. [V]
• Design an enterprise RAG pipeline on vector databases for unstructured data. [V] [~]
• Design RAG over confidential documents with strict per-user access control. [V] [~]
• Debug a cascading failure where an upstream agent passes malformed context downstream. [V] [~]
• Build an automated evaluation pipeline for regression, bias and hallucination in a summarisation tool. [V]
• Build a real-time validator that blocks hallucinated or out-of-schema output. [V] [~]
• Plan a RAG project and discuss the model lifecycle. [R]

Agentic-AI questions

• Design an agent that searches and synthesises unstructured internal documents with RAG. [V]
• Design an internal-search agent with accurate access control, no hallucination and real-time index updates. [V]
• Architect a story-writing agent with long-range consistency, character memory and state tracking. [V]
• Design a BigQuery agent that turns natural language into optimised multi-stage pipelines. [V]
• Define tools, function calling and recovery for unexpected API failures. [V]
• Architect an agentic cloud-security system that detects, analyses and mitigates incidents in real time. [V]
• Design a multimodal smart-glasses agent that plans actions from camera and speech. [V]
• Handle tool-calling failures and non-deterministic loops in production. [V]
• Design a multi-agent system that analyses security logs, isolates compromised assets and writes incident reports. [V]
• Build an evaluation suite for trajectory accuracy, tool-selection precision and hallucination rate. [V]
• Build a benchmark checking whether an analytics agent chose the right SQL tools and schema parameters. [V]
• Task success drops after a foundation-model checkpoint update. What do you do? [V]
• Design infrastructure for an enterprise agent serving millions of concurrent requests. [V]
• Manage and invalidate context caches over dynamic documents. [V]
• Implement a thread-safe conversation buffer with token truncation. [V]
• Given tool dependencies, detect cycles and output a valid order. [V]

LLM serving and ML infrastructure

• Design a distributed LLM serving engine on TPU/GPU clusters: KV-cache compression, PagedAttention, continuous batching, multi-region balancing. [V]
• Serve an ads LLM at 100,000 QPS with p99 under 100 ms. [V]
• Design a Spotify- or Google-Play-style real-time recommender across candidate generation, ranking and re-ranking. [V]
• Combine real-time feature retrieval with batch deep-learning inference in a music recommender. [V]
• Add an LLM agent for natural-language follow-ups to a music recommender. [V]
• Design a multi-stage recommender with production monitoring. [V]
• Cut latency and compute when deploying multimodal generative models to edge devices. [V]
• Handle cold start and real-time context in large recommenders. [V]
• Route queries between lightweight and frontier models under cost budgets. [V] [~]
• Reduce latency and token cost without hurting quality. [V] [~]
• Serve multimodal Gemini models across distributed edge nodes. [V] [~]

ML theory and evaluation

• Derive cross-entropy and hinge loss, with gradients and loss-landscape properties. [V]
• Design an evaluation framework for a production generative app; contrast ROUGE/BLEU with LLM-as-judge and human review. [V]
• Compare LoRA and QLoRA with full fine-tuning: theory and memory savings. [V]
• Trade-offs among cross-entropy, triplet loss, and DPO for tool-selection fine-tuning. [V]
• Explain attention and how context compression affects retrieval accuracy. [V]
• Which metrics monitor drift and degradation in production agents? [V]
• Design continuous evaluation for trajectory accuracy, tool selection, and hallucination. [V]
• How do loss functions differ for free text versus structured tool calls? [V]

Coding questions

Agent- and retrieval-flavored problems

• Given N CPUs and M task durations, find the minimum completion time. [V]
• Count valid unlock patterns of length at least L on an N x M grid with eight-direction moves and no jumping over unvisited dots. [V]
• Find the most frequent event and its timestamps; extend to a sliding window. [V]
• Build a balanced search tree or graph from a sorted array under structural constraints. [V]
• Topologically sort subtasks; traverse an N-ary tree in place. [V]
• Weighted directed graph: can the source reach the target, and what is the shortest path? [V]
• Implement a QuadTree with insert, point, and range queries. [V]
• DP: choose the action sequence maximizing cumulative reward under state constraints. [V]
• Thread-safe conversation buffer with token truncation. [V]
• Detect circular tool dependencies and return a valid order. [V]
• Embedding cache to avoid redundant vector searches. [V] [~]
• Parse streaming LLM output and enforce a JSON schema. [V] [~]
• Dynamic chunking and token distribution for long documents. [V] [~]
• Thread-safe state manager for agent execution history. [V] [~]

Additional reported problems

L6 report

• Logger or timer with added constraints
• Grid problem from O(n^2) to O(n)

Vendor-tagged

• Substring search with weighted sampling
• Streaming weighted sampler
• Connected crop layout and safe paths
• Fence-painting DP
• Robust Python generator
• LRU-cache variant
• Trie autocomplete
• Multi-tenant rate limiter

Reported

• Maximum path sum in a binary tree
• Decode string
• Word ladder
• Longest increasing path in a matrix
• Snapshot array
• Minimum window substring
• Maximal rectangle

Interview Query list

• String shift
• Bigrams
• Bracket matching
• 2D rain water
• Nearest common ancestor
• Minimum change
• Basic regex (. and *)
• First to roll a six
• 500 cards, three draws in increasing order
• Explain a p-value to a layperson

Product, behavioral, and responsible AI

AI-specific

• Choose MVP functionality for a complex enterprise GenAI case. [V] [~]
• How do you decide a prototype is ready for production? [V] [~]
• Evaluate cost per request, time to first token and tokens per second against UX SLAs. [V] [~]
• A feature has high engagement but unacceptable hallucination or safety risk. What now? [V] [~]
• Turn a fragile GenAI prototype into a fault-tolerant product. [V] [~]
• Balance accuracy against latency under a tight deadline. [V]
• Resolve conflict between research recommendations and infrastructure limits. [V]
• Describe a deployed AI system that failed; diagnosis and prevention. [V]
• A collaborator took public credit for your work. [V]
• You had several P0 tasks. How did you realign stakeholders? [V]

Googleyness and general

• Conflict, failure and proudest-project stories using STAR. [V]
• What hurdles have you hit in data projects? [R]
• How do you handle conflict and ambiguity? [V]
• Present and defend your research and explain your interests (intern). [V]
• Why Google? Favourite Google product? Conflict on your team? [R]
• How do you approach GenAI safety in consumer products? [P]

---

Cross-company LLM, RAG, and agent questions

• Single-GPU batching for up to 100 synchronous inputs. [P]
• End-to-end LLM batching; efficient sampling API. [P]
• Insurance-claims agent with RAG and token-cost control. [P]
• Customer-service agents (outdoors retailer; streaming service). [P]
• The model is confidently wrong in high-risk settings. [P]
• Explain RAG; prompting vs RAG vs fine-tuning. [P]
• LLM-as-judge failure modes; catching regressions on a model swap. [P]
• Async retries with timeouts and exponential backoff. [P]
• Detect circular references among spreadsheet cells. [P]
• In-memory key-value store with begin, commit and abort. [P]
• RAG-based agent; agent harness and evaluation system. [P]
• Distribute model weights to a GPU fleet. [P]

Practice drills (our synthesis, no interview evidence)

• Estimate GPU memory to serve a 70B-parameter model at FP16 versus INT8 with 32 concurrent 8K-token requests; separate weights from KV cache. [P]
• When does speculative decoding help, and when does it hurt? [P]
• Compare post-training quantisation with quantisation-aware training; which accuracy risks do you test? [P]
• Implement reciprocal rank fusion for hybrid retrieval and explain tuning. [P]
• How can a retrieved document carry a prompt injection? Defences at each stage? [P]
• When would you use MCP versus A2A? [P]
• Design an orchestrator with sub-agents; mark idempotent calls, approval gates and rollback. [P]
• Design invalidation for a semantic cache over live documents. [P]
• Validate an LLM judge against human labels before using it in CI. [P]
• Triage a task-success drop after a checkpoint update. [P]

Full guide: Link

AI Data Companies: Thoughts on Org / Comp Structure

I've been thinking about the right org structure and compensation philosophy for AI data companies.

Rough thoughts: (THIS IS NOT LEGAL ADVICE)

1. Avoid incorporating as a C-Corp due to double taxation of dividends. Instead, create a S-Corp or LLC.

2. Don't raise venture money, at least initially. Raise money from family offices or angels since VCs can't invest in a S-Corp or a LLC. If the company does turn out to have durable equity value, you can always switch to a C-Corp (S-Corp might be better for this reason than LLC, but pls consult your lawyer).

3. Raise the amount (+ buffer) you need to get to cash flow positive (I think this is likely sub-5M, probably even 1-2M). Obviously this is easier said than done for most, but if you have a strong background and a differentiated POV + lab connects, there are TONS of folks who'd be interested in this kind of structure.

4. (this will be controversial) Don't give any equity to employees. Everyone should be paid a relatively standard base + a highly variable bonus, which is a % of gross profit they are tied to, similar to a hedge fund.

5. After all employees are paid out, keep 30% of the remaining cash on the balance sheet and dividend out the rest to shareholders.

Run the company like a multi-manager hedge fund. (a) small P&L-owning pods per lab or vertical (b) a thin shared platform (c) an independent QA function with veto power, since QA is SUPER important (d) a tiny full-time core, and (e) an elastic expert network of contractors.

In general, I think a hedge fund style org and comp structure is the best way to go here. Most hedge fund employees don't own equity in the underlying management company, but nobody ever complained of not owning equity in Jane Street. AI data companies are similar to hedge funds because:

(a) The edge decays. A hedge fund’s alpha erodes as others find the same trade. A data company’s edge erodes too: today’s frontier data (a new domain, a new eval format, a new kind of expert annotation) becomes commoditized within a year or two.

(b) Returns come from people, and you can attribute them. At a fund you know which PM or desk generated the P&L. At a data company you know which team owns which lab relationship, which project, which data vertical. When contribution is measurable, a cut of gross profit is a fairer and more motivating than a slice of a cap table whose value nobody can predict.

(c) It doesn’t need much capital. Like a fund’s management company, a data business is mostly people and some tooling. Once it’s cash-flow positive, more capital doesn’t make it grow much faster.

If the business looks like a high-margin, people-driven, cash-generating machine with uncertain long-term moat, you should structure it like one. Optimize for distributable cash and paying the people who generate it, and keep the option to become a “real” equity story if one emerges.

Weight gain and loss : A reality check

Can 3 months change years of habits? So the two types of weight gain that we have had, one is our essential or habitual weight gain, which has happened because of a change of habits from being very active as youngsters to now being relatively sedentary as adults. The second part is purely because of different circumstances that we have gone through.

Now, the approaches for both are drastically different. The circumstantial one goes off by offsetting that high calorie intake in bursts of short duration. Like you do a three-month keto, for example, you start exercising hard, you’ll offset it. But what about the essential weight gain?

That will not go with these short-term approaches. That requires strategic thinking over a period of two to three years. As you systematically drop your baseline calorie intake, up your baseline physical activity, deal with stress well, not use food as a coping mechanism, sleep better, take your supplements.

All this over a period of three to five years will systematically bring your baseline weight down. My suggestion to you, if you have experienced both sets of weight gain, don’t stop after the first half of your weight loss. The real game begins for the second half, which will test your habits and the way you look at your body and lifestyle in general.

AI and Data Control

AI can infer your risk of dying from your face but Who controls what it does with your data?

Imagine your friend has just been diagnosed with cancer. The doctor goes over the treatment options, but everyone is still worried about one thing: the risk of dying. Then, someone you know takes a photo of your friend's face. In just a few seconds, an AI system estimates your friend's risk of dying in the next three months. It's hard to believe and feels shocking.

But researchers have created an AI system called FAHR-FaceSurvival. It looks at facial photos to estimate survival risk for cancer patients. The model isn't perfect, but it did find that people with higher risk scores were more likely to die. Nevertheless, these findings are remarkable but breakthroughs in image recognition are challenging as you can see from last week’s stories.

• One company used AI to make food photos that didn’t match what customers actually got.
• An artist found that an AI-generated video seemed to use her identity without asking.
• An investigation showed how a personal AI assistant could build detailed profiles about people’s relationships, preferences, and even their vulnerabilities.

Let’s take a closer look at what happened.

1. AI turns breakfast into a glass of cola with bacon and eggs floating in cola

Just Eat Takeaway.com used AI to create menu images for restaurants, but the technology sometimes got dishes wrong.

For example, one cafe's sausage soda bread breakfast dish was illustrated as a pint of cola with meat floating in it. Just Eat apologized, took down the inaccurate images, and paused the rollout.

AI-generated images can look professional, polished and convincing. But that doesn't mean they accurately represent the product, service, or information customers rely on.

2. When AI uses someone's identity without consent

A Manchester-based musician says an AI-generated video that appears to feature her likeness was shared on Israeli government social media accounts. The artist publicly objected, saying she had not consented to the use of her identity or image. The posts were subsequently deleted.

The leadership question is no longer just, "Can AI create this?" It must also be:

“Do we have the right to use this data, and have the people involved given their consent?”

What Organizations are doing: Establishing boundaries for AI-generated content

Vinted, a resale marketplace, shows how companies can set clear rules for AI-generated content because users of the second-hand seller app Vinted have recently been discussing a seeming uptick in the use of AI-generated photos to sell products. The company has made it clear that using AI is not always banned. For instance, sellers can use AI for things like removing backgrounds from photos. But sellers still need to use real photos that show the actual items for sale.

AI cannot be used to make an item look different from its real appearance, condition, or authenticity. Vinted also uses moderation teams, member feedback, and detection tools to spot misleading content. If rules are broken, they may remove listings or suspend accounts. Instead of pausing or banning AI or letting people use it however they want, companies can set clear rules about what is allowed and make sure those rules are followed.

This week's lesson for Leaders

Good AI governance begins by telling the difference between:

• What AI is technically able to do?
• What the company should let it do?
• What data the system is using with permission?
• What checks and systems the company has to verify and protect its processes?

The Bottomline

Clear principles and governance don't always limit new ideas. They differentiate an organization's successes and failures.

AI is not contributing to productivity

Companies are spending millions on AI. New research suggests it's NOT paying off. And the reason might have nothing to do with the technology. A recent analysis from Apollo's Chief Economist, Torsten Slok, raises a troubling question about AI's financial payoff.

While the seven largest tech companies are seeing revenue per employee and profit margins rise, those measures have remained largely flat across the other 493 companies in the S&P 500. That's a pretty big warning sign. But here's what I think we're missing.

Companies are investing in AI models, licenses, and tools without putting nearly enough thought into what happens next. They expect productivity gains simply because employees now have access to AI. But access isn't adoption. And adoption doesn't automatically create business value.

You still need:

• A clear business problem worth solving.

• Processes redesigned around how work actually gets done.

• Employees who understand how and when to use AI.

• Leaders who can manage the changes.

• A way to measure whether any of it is working.

Otherwise, you're adding another technology expense without changing the business. And that's what worries me. We keep talking about how powerful AI is. Maybe we should be asking whether organizations are prepared to use that power. Because AI doesn't transform a business. People do.

What do you think? Are companies investing too much in AI tools and not enough in the people and processes needed to make them work?

Engineered to Eat More? Let AI Check the Recipe!

Imagine opening a packet of chips intending to eat just a handful, only to find yourself reaching the bottom of the packet minutes later. Or buying a fruit-flavoured drink that looks wholesome, tastes delicious and feels refreshing, without realising how little actual fruit it contains.

Is this simply good food science at work, or has food engineering become so effective that it sometimes works against the consumer?

Food engineering has brought enormous benefits to society. It has improved food safety, extended shelf life, reduced waste, enhanced consistency and made nutritious products more accessible. The problem is not food engineering itself, but how some of these capabilities are used, not just to improve food, but to encourage people to consume more of it.

Some of the most effective food-engineering techniques are not inherently harmful. Carefully balancing sugar, salt, fat, flavour and texture can make products enjoyable and consistent. Creating the perfect crunch, smooth mouthfeel, appealing aroma or satisfying flavour is a genuine technical achievement.

However, these same capabilities can be used to make products exceptionally appealing and easy to consume repeatedly. Highly palatable combinations of ingredients, convenient packaging, large portions and aggressive marketing can encourage people to eat beyond their original intentions.

Think about the snack that seems impossible to put down, the sugary drink that goes down effortlessly, or the ready-to-eat product designed for convenience but offering little nutritional value. Each individual design choice may appear reasonable. Together, they can create an environment in which consuming more becomes easier than consuming mindfully.

This does not mean that every tasty product is deliberately addictive or that every manufacturer intends to encourage overconsumption. Consumer behaviour is complex. Nevertheless, when product development focuses overwhelmingly on repeat purchases, sensory appeal and consumption frequency, nutrition and consumer well-being risk becoming secondary considerations.

Food engineering makes food better. Responsible engineering also considers how much we consume, what we consume and how those choices affect our health.

Can AI become a force for responsible food engineering?

Artificial intelligence offers an opportunity to bring greater accountability into food development and manufacturing.

Imagine an AI system that analyses a product's ingredients, nutritional composition, processing methods, portion sizes and marketing claims. It could flag excessive levels of added sugar, sodium or saturated fat, identify potentially misleading health-related claims, and highlight formulations that may warrant further assessment because of their nutritional profile and intended consumption patterns.

AI could also help manufacturers reformulate products by finding ways to reduce sugar, salt and unhealthy fats without sacrificing affordability, food safety or consumer acceptance. By evaluating thousands of ingredient combinations, it could accelerate product development and reduce the trial-and-error process traditionally used by food scientists.

More importantly, AI could help manufacturers look beyond the question, “How do we make consumers want more?” and ask, “How do we make a product people enjoy while supporting healthier consumption?”

Such systems must be designed carefully. AI cannot reliably determine whether a food is intentionally engineered to encourage overconsumption simply by analysing its ingredients. It would need relevant nutritional standards, validated research, reliable product data and appropriate assessment of portion sizes and eating behaviour. Independent testing and human oversight would remain essential.

The objective is not to dictate what people should eat. It is to give them better products, clearer information and a fairer opportunity to make informed choices.

AI is already demonstrating its value in food manufacturing through quality inspection. One reported industry case involving an Indian potato chip producer described substantial product losses caused by quality defects and limitations in manual inspection.

The manufacturer introduced AI-powered visual inspection to identify undesirable potatoes before processing. According to the published case study, the system was more than 90% effective at identifying undesirable potatoes and detected 15% more defects than manual inspection methods.

The challenge was inconsistent defect detection and the resulting quality losses. The solution was to use AI to inspect products more consistently and identify defects that human inspection could miss.

Although this example addresses quality control rather than overconsumption or misleading nutritional formulation, it illustrates an important principle: AI can make food manufacturing more measurable, consistent and accountable. Similar capabilities could be combined with nutritional analysis, recipe monitoring and label verification to support a more comprehensive approach to responsible food production.

The opportunity now is to extend AI beyond identifying defective potatoes and towards evaluating whether the finished product meets nutritional commitments, accurately represents its ingredients and supports responsible consumption.

Food manufacturers operate in competitive markets, and taste, convenience, affordability and profitability will always matter. The goal is not to eliminate commercial success or make food less enjoyable. It is to ensure that commercial success does not depend on encouraging excessive consumption at the expense of consumer well-being.

Manufacturers can use AI to improve nutritional quality, optimise reasonable portion sizes, reduce waste and develop appealing products that balance enjoyment with health. Regulators can establish clearer standards for nutritional claims and marketing practices, while independent audits can help ensure that AI systems deliver what they promise.

Ultimately, AI is a tool, not a moral compass. If it is programmed only to maximise sales and repeat purchases, it may make existing problems worse. If it is designed to balance commercial objectives with nutrition, transparency and consumer welfare, it could become a powerful ally.

We should be able to enjoy our favourite snacks without needing extraordinary willpower to stop eating them, or a detective's skills to understand their labels.

The real opportunity is not to stop food engineering, but to stop rewarding engineering that encourages consumption without considering its consequences.

Perhaps the most meaningful measure of innovation in the food industry should no longer be how successfully a product persuades consumers to eat more, but how successfully it delivers taste, convenience, nutrition and trust.

That is where AI can make a difference: helping the industry engineer food that people love, without losing sight of the people eating it.

#ArtificialIntelligence #FoodEngineering #FoodTechnology #FoodManufacturing #ResponsibleAI #Nutrition #FoodSafety #ConsumerAwareness #FoodInnovation

Strength Training - Temptation

Strength training has a way of tempting you. Quite a few people ask me how to begin. And truth be told, it’s often what happens after they begin that I want to talk about. A few sessions go well, the weight feels a little lighter, and suddenly the next dumbbell looks quite inviting. You weren’t planning to rush anything. But now that you’re here, why not lift a little more?

I understand the enthusiasm. It feels good to discover that something which was difficult is getting easier. But before you increase the weight, notice how you’re finishing those repetitions. Are you still controlling the movement, or has the rest of your body quietly volunteered to help with that one bicep curl?

That’s where I’d ask you to take your time. Especially if you’re coming back after a long break. There’s a temptation to pick up where you left off, as though your body has been patiently maintaining everything while you were away managing work, family and life. Start lighter if you need to. There’s nothing to make up for.

I’ve put together two charts with six exercises each for your upper and lower body. Learn the movements from a trainer first, then practice with a weight you can control for 12–15 repetitions. Build up gradually, paying attention to your technique and recovery. The thing I keep coming back to is this - I want strength training to become something you comfortably make room for in your life.

Some sessions, doing your exercises well and coming back the next week is plenty to feel good about.