When I speak at conferences, people often tell me I did a great job. My first thought, before I even finish hearing the sentence, is that they're being polite.
‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ 

Click here to read this mailing online.

Your email updates, powered by FeedBlitz

 
Here is a sample subscription for you. Click here to start your FREE subscription


  1. The Politeness Excuse.
  2. The Human Review Step Is Not Optional. Three Companies Proved It This Week.
  3. The Reunion the Algorithm Would Never Have Found
  4. The Radio, the Phone Number, and Hearing Yourself on the BBC
  5. The Antidote to Uncertainty: Why the Capacity to Learn is Your Ultimate Safety Net
  6. More Recent Articles

The Politeness Excuse.

When I speak at conferences, people often tell me I did a great job. My first thought, before I even finish hearing the sentence, is that they're being polite.

I've decided that thought is wrong, and I want to explain why, because I think a lot of us are carrying the same wrong thought without examining it.

Growing up, the warning came early and often. Don't show off. Don't let the neighbors see too much. If your son did well in exams, you mentioned it once, in passing, and changed the subject. The logic underneath it was never spoken directly, but everyone understood it: visibility invites envy. Success was something to manage quietly, almost apologetically, like a small fire you didn't want the wind to find.

That warning was never universal, not even within the same country. In other parts of India, I watched families spend on weddings and celebrations that had nothing to do with what they could afford and everything to do with what they wanted people to see. Bigger tents. Louder bands. Nobody was hiding from the neighbors. The neighbors were the point.

Two completely different relationships to being seen, and I've come around to thinking the second household had something right that mine didn't. Not the spending. The refusal to apologize for being visible.

Here's what I mean when I say I did a great job speaking. It took years of speaking badly before I learned to speak well. It took recording myself, watching the playback, cutting the filler words, rebuilding the pacing one talk at a time, for longer than I'd like to admit. When someone tells me that work showed up on stage, the honest response is to take it. Not loudly. Not with a speech about the journey. Just a clean, simple yes, you're right, I worked at this.

That's the part I want to get better at. Owning it out loud. Letting a compliment land instead of bouncing it back across the room like a hot plate. The work was real, the hours were real, and pretending otherwise doesn't make me humble. It just makes the work invisible, which is a strange thing to do to something I'm proud of.

So here's where I've landed, and where I'd like more of us to land. Share the win. Say thank you and mean it. Let people see that the practice paid off, because somewhere in that audience is someone who needed proof that the unglamorous repetition actually leads somewhere.

The guilt, if there's room for any, belongs somewhere else entirely. Not in being seen doing well. Save it for the times you showed up, did the work, and it still didn't land the way you hoped. That's a real feeling, earned and specific, and it has nothing to do with a compliment after a good talk. Confusing the two, treating every win like something to apologize for, just because effort doesn't always come with a guaranteed outcome, is its own kind of waste.

You probably have your own version of the stage. A pitch you finally nailed after the third rewrite. A kid you raised who turned out fine. A business that took ten quiet years before anyone called it an overnight success. Somewhere in there is a compliment you waved off this week, and a parent's old voice telling you not to let the neighbors see. You can put that voice down too. It was never protecting you from envy. It was just protecting you from being seen, and those are not the same thing.

I'm trying to retire the old reflex. Next time someone tells me I did a great job, I want my first thought to be thank you, not they're just being polite. If you're carrying the same reflex, consider this your permission to retire it too. That shift, small as it sounds, is the whole essay.

   

The Human Review Step Is Not Optional. Three Companies Proved It This Week.

Analysis · Enterprise AI · June 2026

Three companies released data about AI and human oversight in the same week. None of them coordinated. All three pointed in the same direction.

>80%
Anthropic code
written by Claude
(May 2026, company-supplied)
75%
Google code
AI-generated
(Cloud Next 2026, company-supplied)
Engineer code
output increase
at Anthropic vs. 2024

Key Takeaway

Anthropic and Google independently disclosed that AI now generates the majority of their own production code. Cisco staged a demo showing what happens when you remove review from the loop. The pattern across all three is the same: more automation requires more oversight, not less.

Sundar Pichai put the number in a blog post during Google's Cloud Next 2026 conference on April 22. Seventy-five percent of all new code at Google is now AI-generated and approved by engineers, up from 50% the previous fall. It appeared as a productivity proof point, not a warning. The 75% figure arrived as good news.

Six weeks later at Cisco Live 2026, I watched Jeetu Patel open his Day 2 keynote with a customer service scenario. A refund agent, operating without oversight, had paid out $1,200 to a customer who had paid $700. The answer Cisco demonstrated was not to remove the agent but to add a review layer. Galileo, which Cisco acquired specifically for agent observability, flagged the transaction, held the payout, and routed the discrepancy to a human for approval before it completed. The $500 error never went out.

Then on June 4, Anthropic published "When AI Builds Itself." More than 80% of the code in Anthropic's own production codebase is now written by Claude. Not assisted. Written. Engineers merge roughly eight times as much code per day as in 2024. The report called for a coordinated pause among frontier artificial intelligence labs before development crosses into recursive self-improvement, the point at which an AI system can design its own successor without human intervention.

On June 4, Jack Clark, one of Anthropic's co-founders, gave his own BBC World Service interview about what the data meant. Nigel Doran had originally emailed about Anthropic's IPO filing, but Clark's interview that day shifted what we actually talked about. By the time Krupa Padhy introduced me on the weekend program on June 7, the IPO was the secondary story. My answer was what I would tell any enterprise technology leader: the message is not that AI should stop. Every output needs a human to review and approve it before it goes further.

The IPO Framing Is Worth Taking Seriously Before Dismissing It

The prevailing read on Anthropic's pause call is that it is credibility-building timed to a capital raise. Anthropic filed confidentially for its IPO one week before publishing this report.

What it cannot account for is the data Anthropic chose to disclose alongside the argument. More than 80% of merged production code written by Claude. Engineers outputting eight times the code of 2024. Leadership estimating the real figure may exceed 90% including scripts and experimental work. These figures complicate a public offering. They raise questions about auditability and what happens to valuations if AI-written code fails at scale. A company does not file for an IPO and then volunteer data that makes institutional investors nervous unless the concern is genuine.

And Pichai's April post came with no comparable motive. Google has no IPO pending, no safety positioning to establish. The 75% figure was productivity evidence. Two of the most sophisticated engineering organizations on earth disclosed the same structural shift from opposite ends of the motivation spectrum, six weeks apart. That convergence is the evidence.

The Network Demo Shows What the Solution Looks Like

The second scenario Cisco demonstrated was a network operations failure. An IT monitoring agent detected the fault, traced it to root cause, and worked out the fix. Then it stopped. It sent the engineer its full reasoning through WebEx and waited for approval before acting. The engineer reviewed it, agreed, and the network recovered in 45 seconds. After that, the system offered a standing choice: for this class of problem going forward, should the agent act without waiting? The engineer decides. Trust extends one verified decision at a time.

The observability layer that makes this practical at enterprise scale is Galileo, acquired by Cisco, whose Luna model routes all agent evaluations through Cisco's own systems at 98% lower cost than conventional monitoring, according to Cisco's figures. Splunk, now part of Cisco, provides the unified view across network, application, and AI system data that made the 45-second diagnosis possible.

The message is not that AI should stop. A human being needs to review and approve the work before it goes further. That sounds obvious. It is apparently still worth saying out loud.

Pichai's disclosure carries the same logic. "AI-generated and approved by engineers." He put the approval clause in the same sentence as the percentage. Whether enterprise leaders treat that approval step as essential infrastructure or as friction to eliminate is the operational question Cisco's demos were designed to answer.

The Jobs Question Has a Longer Answer Than Most Coverage Gives It

The leaders worth listening to are not framing this as replacement. They are asking what their teams can accomplish now that was not possible before. That is a different question, and it tends to produce different answers than the ones dominating the public debate about AI and jobs.

Some roles will evolve. Some will disappear, as there are no longer operators running elevator cars or flagmen walking ahead of automobiles. That has always been true of technology. The difference now is speed, and whether the transition is managed or damaging depends on whether a human is still in the room when the consequential decisions get made.

Key Takeaway

Anthropic's IPO timing is a legitimate lens for reading the pause call. It does not explain why they disclosed figures that complicate their own valuation story, or why Google corroborated the same trajectory six weeks earlier with no comparable incentive. Human approval is not optional overhead at this production percentage.

CIO / CTO Viability Question

If more than 75% of code at the most advanced AI engineering organizations is now AI-generated, what percentage of your deployed agents' decisions are reviewed by a human before they execute, and what does that answer look like at ten times current transaction volume?

Sources

Pichai, Sundar. "Sundar Pichai Shares News from Google Cloud Next 2026." Google Blog, Google, 22 Apr. 2026, blog.google.

Favaro, Marina, and Jack Clark. "When AI Builds Itself." Anthropic Institute, Anthropic, 4 June 2026, anthropic.com.

Patel, Jeetu. Day 2 Keynote Session. Cisco Live 2026, Mandalay Bay Convention Center, Las Vegas, 3 June 2026.

Doran, Nigel, and Krupa Padhy. Technology Interview. BBC World Service Weekend Programme, BBC, 7 June 2026, bbc.co.uk.

Bellamkonda, Shashi. "Cisco Live 2026: When AI Agents Go Wrong." shashi.co, 3 June 2026, shashi.co.

Bellamkonda, Shashi. "Anthropic's Platform Bet: Code with Claude 2026 Was Not a Product Launch." shashi.co, 7 May 2026, shashi.co.

   

The Reunion the Algorithm Would Never Have Found

Saturday afternoon, nothing on Netflix worth clicking, Prime suggesting the same rotation of prestige drama I'd already half-watched. I went looking instead. That's how I found myself watching Patriot, a Malayalam political thriller from 2026, on my couch in the middle of the day, in the original Malayalam with English subtitles, following two of Indian cinema's greatest actors through a surveillance conspiracy that felt uncomfortably plausible.

I grew up with Malayalam cinema in the background. My parents' generation watched Mammootty and Mohanlal the way Americans followed Pacino and De Niro, with that same mixture of admiration and proprietary pride. These were our actors. The kind of performers whose faces alone carried meaning before a word was spoken. I drifted, as second-generation kids do, toward whatever the mainstream algorithm served. Hollywood took up the space.

What's changed is not my taste. What's changed is the door.

Streaming platforms and subtitle technology have done something quietly remarkable. They've made world cinema genuinely accessible in a way that feels different from the old subtitled art-house experience, which always carried a slight self-consciousness, a sense that you were doing something deliberate and worthy. Now you can simply watch. The friction has come down enough that curiosity is sufficient. You don't need to be a devotee. You just need an afternoon and the willingness to read.

There is something that happens when actors of this caliber share a scene together, something that has nothing to do with the language they're speaking or the language you're hearing. It crosses. It has always crossed.

And what an afternoon to stumble into. Patriot reunites Mammootty and Mohanlal on screen for the first time in over thirteen years. Directed by Mahesh Narayanan, it is a grounded spy thriller, no slow-motion fan service, no mass masala theatrics, just two older men playing characters who carry weight and history in their postures. Mammootty plays Dr. Daniel James, a senior government scientist falsely branded a traitor. Mohanlal arrives later as Colonel Rahim Naik, his trusted old friend, and the film shifts when he does. There is a sequence where the two communicate in Morse code to evade the surveillance systems hunting them, and it is the most quietly thrilling thing I've seen in months. Old-school craft defeating high-tech paranoia. It felt like something the actors understood personally.

Fahadh Faasil plays the antagonist Shakthi, and he is the kind of villain who doesn't need to raise his voice. One of the most compelling performers working in Indian cinema right now, he brings a cold, coiled menace to the role that stays with you after the film ends. There is a particular kind of screen presence that communicates threat without theatrics. Faasil has it in abundance. The ensemble around these three is extraordinary as well, Kunchacko Boban, Nayanthara, Revathi, names that mean something to anyone who grew up watching Malayalam films and that deserve to mean something to everyone else.

The film has flaws. Some of the plotting telegraphs itself. A few prominent roles, particularly among the women, feel underlived given the talent. But the craft is there: cinematographer Manush Nandan gives the film a visual register that could hold its own against any international thriller, and composer Sushin Shyam resists every temptation toward bombast, letting silence do the work that lesser films bury in noise.

What I keep returning to is not the plot but the faces. There is something that happens when actors of this caliber share a scene together, something that has nothing to do with the language they're speaking or the language you're hearing. It crosses. It has always crossed. The willingness to read subtitles is a small thing to ask for that kind of payoff.

I used to think of watching Indian cinema as going back. Back to something rooted, something connected to where my family came from. Patriot made me reconsider that framing. This is not backward-looking cinema. It is deeply engaged with the present: with surveillance states, data privacy, political dissent, the fragility of institutional trust. It happens to be in Malayalam. It happens to star two men who have been making films longer than I've been aware of films. None of that makes it a relic. It makes it richer.

There is a whole world of cinema beyond the three platforms and the one dominant industry. Some of it is extraordinary. The algorithm will never serve it to you. You have to go looking. Saturday afternoons are good for that.

   

The Radio, the Phone Number, and Hearing Yourself on the BBC

The summer I turned ten in Madras, the BBC was already a presence in our house. Chennai is what the city is called now, but in my memory it is always Madras, and Madras in summer is heat and humidity and the particular sound of a valve radio warming up in a room where the ceiling fan turns slowly overhead. My parents listened to the BBC World Service the way other families might have kept a clock running. It was always on. It marked time. It told us there was a wider world out there, even when that world felt very far away.

I never imagined I would be in it.

This past week, Krupa Padhy of BBC Weekend introduced me to her audience as Principal Research Director at Info-Tech Research Group and asked me to respond to something Jack Clark, one of Anthropic's co-founders, had told the BBC the day before. Clark had spoken about research showing that eighty percent of Anthropic's own coding is now done by AI, and about why that capability should not be allowed to run without humans in the loop. My answer was straightforward: the message is not that AI should stop, but that a human being needs to review and approve the work before it goes further. That sounds obvious. It is apparently still worth saying out loud.

What I was not prepared for was the particular strangeness of hearing myself come through the same channel that once brought the world into a house in Madras. My old Siemens valve radio is long gone. I listen to the BBC now as a podcast or through their app on my phone. But something about that moment felt continuous with the ten-year-old who sat in a hot room and learned that voices could travel across oceans.

The first international call I ever made in my life was to BBC Bush House in London. It was the only international phone number I had.

My friend Ajoy Joshi heard the interview from Australia. He sent me a message because he could not quite believe his ears. That is one of the great gifts of a moment like this: it becomes real when someone who knew you before any of this hears you and recognizes you anyway.

This week held other things worth pausing on. Mike Isaac quoted me in the New York Times in his piece on Anthropic's IPO filing. And Jonathan Vanian sought my perspective for his CNBC column on whether Meta can build a business around AI the way it built one around advertising. What I told him was what I genuinely believe: Meta has a long way to go to build an enterprise business from the ground up, because the company has been so focused on consumers. Their Workplace offering always felt half-hearted, because the real attention was on the social side where the ad revenue lives. To compete seriously in enterprise, especially in cloud, you have to build up the processes, the platforms, the technology, and most importantly the people required to actually serve customers over time. Cutting staff while trying to expand in that direction is not a formula that tends to work.

Both conversations were made possible by work I do every day at Info-Tech Research Group, staying close enough to what is actually happening that I can be useful to the 40,000 members who trust us to help them think clearly, either confirming that they are moving in the right direction or giving them something new to push against. I am grateful for that work. It is also what earns the phone calls.

And as if the week needed one more layer, I was at Cisco Live, hearing directly from the leaders of a company that has spent forty years building the infrastructure underlying how the world communicates. There is something grounding about being in a room full of people who understand that none of this happens without a network underneath it. Including, in all likelihood, the signal that carried my voice from a BBC studio to Ajoy's phone in Australia.

Some weeks remind you that the connections you make are not accidental. The valve radio and the podcast app. The one international phone number I memorized as a child and the BBC interview I never planned for. The forty-year-old networking company and the researcher who is still, after all this time, just trying to keep learning.

I have been very lucky. I know that. I want to say it plainly rather than dress it up.

Shashi Bellamkonda is Principal Research Director at Info-Tech Research Group, with affiliations at Georgetown and Stony Brook. He writes at readythoughts.com and shashi.co.

   

The Antidote to Uncertainty: Why the Capacity to Learn is Your Ultimate Safety Net

Growing up in India, the word “exam” carried a heavy, almost suffocating weight. The 90 minutes you spent at a desk frantically writing on paper took on life-changing proportions. The pressure wasn’t just academic; it felt existential. Tragically, we still see headlines of young lives cut short because of poor marks. The system implicitly taught us a dangerous lie: that a single grade defines the boundaries of your entire life.

If you are carrying anxiety today about a performance review, a missed target, or an uncertain career transition, I want to tell you what the school system didn’t: The snapshot does not define the journey.

When you study people who have sustained success over decades, you quickly realize that very few built a lasting legacy based on their grades. Exams and performance metrics are good for the present moment, but they have zero power over your long-term future.

Success isn’t inherited by the straight-A student; it is claimed by the person who refuses to stop learning.

The Bravery of Starting at Zero

True knowledge isn’t a static certificate you hang on a wall; it’s a living, breathing toolkit. And the most vital tool in that kit is the willingness to repeatedly start over.

Pivoting careers is terrifying. I know this firsthand because my own path has been a continuous medley of reinventions. I remember the distinct vulnerability of standing in front of a computer for the very first time — a Timex Sinclair ZX81, a machine so alien to me that I didn’t even know what I was supposed to type first. There was no manual open in front of me, no one to ask. Just a blinking cursor and the quiet embarrassment of being the slowest person in the room.

I remember the steep learning curves of mastering Lotus 123 and WordStar after that. And then the cycle kept repeating — moving from tech support to program management, product management, social media, PR, executive marketing leadership, and eventually to my current role as an industry analyst. Each shift came with its own version of that ZX81 moment: a new vocabulary I didn’t speak yet, a room full of people who seemed to already know things I didn’t.

The transition I felt most acutely was moving into analyst work. I had decades of practitioner experience, but the analyst’s craft — the structured frameworks, the vendor briefings, the discipline of separating observation from opinion — was genuinely new. There were days I wondered if I had made a mistake, if the credibility I had built in marketing would simply fail to transfer. It didn’t. But I had to let myself be a beginner long enough to find out.

Every single shift required me to humble myself and admit: “I don’t know this yet, but I can learn it.”

If you are feeling the paralysis of a new professional chapter right now, realize that your past expertise isn’t lost — it is simply the foundation for your next layer of growth.

Moving From Panic to Preparation

Anxiety is simply the mind trying to predict a future it cannot see. When facing the unknown, people naturally reach for different anchors — some find solace in prayer, others look for distractions to quiet the noise. But if you want to disarm anxiety at its root, the most powerful thing you can do is prepare.

Preparation is not the same as having all the answers. It is the act of reducing the unknown, one small step at a time. Read the book. Take the course. Test the software. Have the conversation you have been avoiding. Each action shrinks the territory that fear needs to operate in.

When you commit to being a lifelong learner, you reclaim your power. The unknown stops being a threat to your survival and simply becomes a syllabus to be studied. You don’t need to know how the whole movie ends; you just need to prepare for the next scene.

The Only Job Security That Exists

I think back to that ZX81 — the blinking cursor, the blank screen, the complete absence of any instruction on what to do next. I had no idea that learning to navigate that discomfort would become the most transferable skill I ever built.

No single exam, boss, or bad review owns your destiny. Your varied, non-linear experiences are not a distraction from your career — they are your competitive advantage, the depth that no straight-line resume can replicate.

Your capacity to learn is the only true job security that exists. Trust your ability to figure it out. You’ve done it before, and you will do it again.

“The real skill isn’t acquiring knowledge — it’s surviving the gap between not knowing and knowing, without quitting.”

I’d be curious: what was your ZX81 moment — the first time you sat in front of something completely unfamiliar and had to decide whether to walk away or figure it out? I’d love to hear it in the comments.

Image:"Times Square, August 2004. I had no idea what came next. Neither did the billboard."
   

More Recent Articles

You Might Like