About
- I…
Activity
9K followers
-
Tript Singh Lamba shared thisI closed with 𝗠𝗶𝗹𝗲𝘆 𝗖𝘆𝗿𝘂𝘀. On purpose. Stanford University Graduate School of Business had me teach a session in their CPO program this month. Thank you to the faculty and to the room for the hours. The arc I gave them: I have never thought about my career in titles. I have thought about it in hills. Capability growth, not title progression. A hill is a career chapter that forces you to learn something new. It has four parts, and I made the class run all four. -𝗧𝗲𝗿𝗿𝗮𝗶𝗻. 𝗪𝗵𝗮𝘁 𝘄𝗼𝗿𝗹𝗱 𝘄𝗲𝗿𝗲 𝘆𝗼𝘂 𝗶𝗻? The operating model you were actually inside, not the org chart you drew. -𝗧𝗮𝗿𝗴𝗲𝘁. 𝗪𝗵𝗮𝘁 𝘄𝗲𝗿𝗲 𝘆𝗼𝘂 𝘁𝗿𝘆𝗶𝗻𝗴 𝘁𝗼 𝗮𝗰𝗰𝗼𝗺𝗽𝗹𝗶𝘀𝗵? The OKR, or the BHAG that looked unreasonable the day you wrote it. -𝗧𝘄𝗶𝘀𝘁. 𝗪𝗵𝗮𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝗱 𝘁𝗵𝗮𝘁 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝗱 𝘄𝗵𝗮𝘁 𝘆𝗼𝘂 𝗸𝗻𝗲𝘄? Usually the moment the thing you were best at stopped working, the failure or setback. -𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆. 𝗪𝗵𝗮𝘁 𝗱𝗶𝗱 𝗶𝘁 𝘁𝗲𝗮𝗰𝗵 𝘆𝗼𝘂? Stated humbly, because the real ones are usually a little embarrassing. Then the part I believe most, which is not in the framework. The summit is never visible from the bottom. That is the feature, not the flaw. Not knowing whether the top sits at 600 meters or 3,000 is the thing that gets me out of bed. Every hill I have picked has been unfamiliar terrain, with at least one slope on it I had no business being on. And when I do get to the top, I have about three breaths in me. Standing on a summit is the least interesting thing you can do on a mountain. Usually, the moment I can finally see the top, I am already looking at the next hill. 𝗖𝗵𝗼𝗼𝘀𝗲 𝘁𝗵𝗲 𝗵𝗶𝗹𝗹 𝗳𝗼𝗿 𝘄𝗵𝗮𝘁 𝗶𝘁 𝘄𝗶𝗹𝗹 𝗳𝗼𝗿𝗰𝗲 𝘆𝗼𝘂 𝘁𝗼 𝗹𝗲𝗮𝗿𝗻. So the last slide was a 2009 clip of Miley Cyrus singing 𝗧𝗵𝗲 𝗖𝗹𝗶𝗺𝗯. I am aware of how that plays in a room of CPOs. I did it anyway. https://lnkd.in/gzkNhedF Because the whole song is one argument, and it is the same one I had just spent an hour making. It is not about how fast you get there. It is not about what is waiting on the other side. It is the climb. Sixteen years, and we are still writing frameworks to say what she said in a chorus. So here is the question I left them with, and the one I will leave here. 𝙒𝙝𝙖𝙩 𝙞𝙨 𝙮𝙤𝙪𝙧 𝙣𝙚𝙭𝙩 𝙝𝙞𝙡𝙡? 𝓐𝓷𝓭 𝓽𝓱𝓮 𝓱𝓪𝓻𝓭 𝓹𝓪𝓻𝓽: 𝓲𝓼 𝓲𝓽 𝓽𝓸𝓸 𝓼𝓲𝓶𝓲𝓵𝓪𝓻 𝓽𝓸 𝓽𝓱𝓮 𝓵𝓪𝓼𝓽 𝓸𝓷𝓮? 𝓘𝓼 𝓲𝓽 𝓳𝓾𝓼𝓽 𝓽𝓱𝓮 𝓼𝓪𝓶𝓮 𝓬𝓵𝓲𝓶𝓫 𝔀𝓲𝓽𝓱 𝓪 𝓫𝓲𝓰𝓰𝓮𝓻 𝓽𝓲𝓽𝓵𝓮 𝓸𝓷 𝓲𝓽?
-
Tript Singh Lamba shared thisYesterday: the agent, the OS, metasearch, the banks, and loyalty. Five more, no less uncomfortable. 6. Legacy infrastructure is dividing into those who invest their way out and those who cut their way down. Sabre has lost ninety percent of its value and still runs code written before most of its customers were born. Amadeus spends a fifth of its revenue on research and is partnering with Google Cloud. Same industry, opposite bets. History rarely offers a cleaner experiment. 7. Data, not marketing, is now the moat, and almost no incumbent has built for it. Hopper reached $850 million in revenue without spending on advertising, predicting a price at better than ninety five percent accuracy and selling that certainty back to the traveler. A three person team in Austin, Flighty, now forecasts flight delays more accurately than the airlines whose flights they are predicting. 8. The next great travel platform is not being designed in San Francisco. It is already running inside MakeMyTrip and inside Trip.com's WeChat ecosystem, processing trillions in a currency most Western analysts have never had to think about. The West is catching up, five to seven years behind, and calling it strategy. 9. Travel is quietly relabeling itself as financial infrastructure. Ninety percent of Hopper's revenue now comes not from travelers but from the banks that white label it, invisible inside institutions whose names carry more trust than any airline's. Navan went public this year at a valuation north of six billion dollars on the same premise. Travel was never the destination. It was always the entry point into someone's spending. 10. The next great travel company will not resemble a travel company at all. This year, Stripe, Mastercard, and Google quietly built the payment rails for AI agents to transact without a human present. Only two percent of travelers today would trust a machine with full authority over their booking. Four out of five travel executives intend to deploy one anyway. That gap is not a risk. It is the opportunity. I do not manage a fund. I do not want one. I simply notice, on weekends, what other people are paid a great deal to notice on weekdays. Tell me which one of these ten you are already exposed to, and did not know it.Trip.com Official Site | Travel Deals and PromotionsTrip.com Official Site | Travel Deals and Promotions
-
Tript Singh Lamba posted thisI once helped run product for a platform that moved a meaningful share of the world's travelers. That chapter closed years ago. The questions did not. Every autumn the same message finds me, forwarded through group chats, then analysts, then, more than once, someone who manages other people's money. Ten predictions this year, no hedging. Five now, five tomorrow. 1. The travel agent has already been replaced, and the replacement carries no commission. Booking's connected trip transactions grew more than thirty percent this year. In twelve months, search's share of trip research fell from fifty one to thirty six percent, while generative AI's share climbed from six to fifteen. An old habit is dying faster than a new one is forming. 2. The OTA is racing to become an operating system, and may only succeed in becoming a pipe. Booking earned $26.9 billion this year, connected trips growing in the high twenties percent. Above it, Google is building agentic booking directly into search, and OpenAI has already made Expedia a launch partner inside ChatGPT. Holding the customer's attention and keeping the customer turn out to be two different jobs. 3. Metasearch is already dead. Only the share price has not heard yet, Trivago down ninety four percent from its peak. Kayak's own CEO called agentic AI an "existential threat" to his category, then pivoted his pitch to frame Kayak as middleware for the AI agents that will replace it. He did not mean to tell you the category is over. He told you anyway. 4. Banks are sitting on the richest unclaimed vein in the industry, and their own executives are starting to say so out loud. A quarter of all leisure travel spending in America moves through a Chase card. Chase's co-CEO put the gap in six words: "our assets were not differentiated." Capital One heard the same problem and bought Hopper outright rather than confess it on an earnings call. 5. Loyalty, the industry's oldest asset, is being devalued by the very companies that depend on it. American's AAdvantage program is valued at three times American Airlines itself. Delta SkyMiles represents most of Delta's equity value. Both are moving toward dynamic pricing that quietly erodes what every mile in an account is worth. Nobody voted for this. It happens to everyone at once, one repricing at a time. Five more tomorrow. The gap only widens from here.
-
Tript Singh Lamba posted thisFinding the document was never the hard part. Enterprise search is close to solved. Point a good index at everything a company has written and it finds the thing. They work. An index answers where is it. No index answers how this evolved, what was argued, who changed their mind, or whether what we promised happened. The difference is not effort, it is structure. An index returns a document. The second reconciles two records never designed to agree, and there is no document to return, because the answer does not exist until something computes it. Take one. Which promises made on customer calls in the last ninety days have no ticket behind them. No document contains that. The call recording holds one half, the repo holds the other, and the answer is the gap, which nobody ever wrote down. We call it retrieval because retrieval is what we know how to build. It is a join. Nobody wanted it to be, because joins with no shared key are miserable and search demos better. An index tells you what your company has. A warehouse tells you what it knows, including what nobody wrote down. Ask a question that needs two systems to agree and see whether you get an answer or a list of links.
-
Tript Singh Lamba shared thisGit will give you a confident wrong answer. Ask which parts of your system only one person understands and the obvious move is commit counts. An afternoon of work, a clean ranking, usually wrong. Authorship and understanding are different, and only one is easy to count. Whoever wrote ninety four percent of a service may have built it from a spec and moved on. Whoever understands it reviewed most of it, argued the hard parts in a thread, and sat on the calls where customers negotiated the constraints. None of that is in git. A real answer scores four things across four systems. Who wrote it. Who reviewed it with substance, meaning comments, not approvals. Who shows up in Slack when it breaks. Who was on the calls the constraints came from. If that ranking never inverts the commit count ranking, it is measuring typing. Then attach revenue, which nobody does. Not which service is fragile. Which fragile service has eleven enterprise accounts running through it. This degrades on a schedule. AI coding tools write a growing share of commits, making commit count a worse proxy every quarter. partners@productnow.ai Which service breaks hardest if one person leaves this month?
-
Tript Singh Lamba shared thisThe board asked what the AI spend returned. That question is landing in a lot of rooms this year and is close to unanswerable in most of them. Not because finance is slow. The shape is consistent. Three to six AI vendors. Coding assistants, a meeting recorder, something in support, something in the CRM, and at least one thing bought on a card nobody upstream knows about. The spend is legible because it sits on an invoice. The return is not, because it would have to trace through code that shipped, to a customer who asked, to a number that moved. Every vendor will show you a usage dashboard. Seats active, suggestions accepted, hours saved by their own arithmetic. None will grade itself against a business outcome, and it would be strange if it did. That is the structural part. The vendor selling you the tokens cannot be the one who tells you the tokens were worth it. Which is why we do not sell a model. Having no stake in which one you use is not a positioning line. It is what makes the answer credible the day it comes back unflattering. partners@productnow.ai Who at your company would be asked to build that number, and with what?
-
Tript Singh Lamba posted thisAugust 2026. Here is the eighteen month call. By early 2028, context platform will mean two different products, and most buyers will learn the difference the expensive way. The category is forming faster than the definition. Eighty eight percent of IT and data leaders already claim to run one. Ninety three percent expect to treat it as shared infrastructure. A word that arrives that fast arrives empty. Today it means retrieval. A good index over your documents so an agent can find things. Real work, and a problem with a known shape, which is why everyone claims it. The second product is a join. Not finding what exists, but reconciling two systems never designed to agree. What was promised on a call against what sits in a repo. Both answer in plain language and both demo well, which is why they get confused. They separate under load. An index that misses gives a slightly worse answer. A join that misses tells you a commitment was met when it was not. The first wave buys the index, it works, and then someone asks a question needing two systems to agree and gets a shrug. The second wave buys knowing the difference. Quote me in eighteen months. Which one are you buying?
-
Tript Singh Lamba posted thisEighteen months out, fast stops flattering you. "We ship in hours, not sprints" is still a brag today. Not for long. Every org running multiple coding agents against a poly-repo codebase is about to learn this the hard way: the agent isn't slow to understand your customer, it's fast to execute whatever lossy summary of the customer made it into the ticket. A real one, verified: eleven customer calls worth $1.8M in pipeline asked for one specific thing. The ticket said something adjacent and wrong. The agent built the ticket, perfectly, and a $420K account is now staring at escalation, because nobody connected the calls to the code before it shipped. Not a model problem. The information was never missing. It was scattered across five systems that don't talk to each other. Prediction: within eighteen months, "time to ship" stops being the board's metric. "Time to correctly scoped" becomes the one that matters, once a few public, expensive misses make fast-and-wrong look worse than right-on-the-old-timeline. The teams that already joined intent to execution will look prescient. Everyone else will be explaining an escalation to their CFO. Quote me on this one.
-
Tript Singh Lamba shared thisNobody in the building can check that date. A rep gives a customer a ship date. Good faith answer. There is no system anywhere in the company where someone in a revenue seat can verify it. This breaks at a specific size. Two hundred to two thousand people, where sales and engineering stopped sharing meetings two years ago and nobody noticed because the numbers were fine. Poly repo. Three or four AI tools shipping code faster than anyone tracks. Every system is doing its job correctly, which is why it stays invisible. Call recordings capture the promise and cannot see a branch. The CRM holds the account and never had a field for the commitment. The repo holds the real date and no seller has opened it. So the question that surfaces all of it, which promises in the last ninety days have no ticket and no branch, cannot be pointed at any tool in the stack. Three people, two weeks, by hand. Then it goes stale. We built the version that runs continuously. Code and CRM connected, no integration project, first answer on day one. partners@productnow.ai How would your team learn today that a date given last week was wrong?
-
Tript Singh Lamba liked thisTript Singh Lamba liked thisThe Rule of Thirds. Not in photography. In innovation at scale. For a decade, Cisco has been the underdog. Nearly every meaningful win once looked improbable. Oddly, even after a series of successes, we still feel like one. That is an advantage. Cisco once operated more like a holding company. Each business wanted its own P&L, engineering, product, sales and marketing. The career path led toward becoming a GM and running a kingdom. That creates speed. But speed is not velocity. Speed is moving fast. Velocity is moving fast in the same direction. When teams optimize locally, they use different scorecards, stories and sources of advantage. The company moves, but it does not compound. A few years ago, we began changing the model: from a collection of businesses to one platform company. That required a few operating beliefs: 1. Company over team over self. 2. We are builders, not empire builders. 3. Products people love, meaningful adoption and an open ecosystem are the scorecard. 4. Quality is priority zero. 5. Structural advantage matters more than temporary feature differentiation. 6. Think big. Play to win. Build durable growth. 7. Cisco is less like a family and more like a championship team. Everyone earns their spot through urgency, learning, grit and the quality of their ideas. Then came the leadership question: how do you compose a team capable of hyper-innovation at Cisco’s scale? Our answer: the Rule of Thirds. One-third are Cisco insiders. They know how to navigate the machine, understand what makes it powerful and respect the need for reinvention. One-third are outsiders. They bring fresh eyes, systems thinking and competitive instinct. They choose the underdog over the fashionable brand because they want to help define what becomes fashionable next. One-third are entrepreneurs. Founders and former CEOs who have built, failed, adapted and started again. They bring urgency and ownership while learning to turn entrepreneurial instinct into scaled systems. All three need one thing in common: ambition unconstrained by conventional wisdom. They are allergic to mediocrity. They take business performance personally. And they care more about configuring the team to win than playing for individual stats. FY26 was the first year the market began to see the results of this platform model. But one good year proves very little. The real test is sustaining the resolve for decades, through inevitable setbacks. A company, like a product, is always in one of two states: incomplete or obsolete. So we will always be a work in progress. We owe that to our customers. Our partners. And the people who keep getting 1.27% better every day because they understand the power of compounding. For anyone entering the workforce: shortcuts exist, but they are usually short-lived. Sustained persistence beats short sprints. If this excites you, find a way to come build at Cisco. If it doesn’t, choose something that does. YOLO. Make it count.
-
Tript Singh Lamba liked thisTript Singh Lamba liked thisToday, I was published in The Wall Street Journal. And honestly, it feels incredible :) I’ve spent the past few years creating personal finance content on social media, including a lot of content about credit cards. A couple of weeks ago, I received an email from Oyin A. , who was working on an article about high-end credit cards and how changing rules are making it harder to maximize their benefits and perks. She wanted to interview me for the piece. I said yes, and the rest is history. Thank you so much to Oyin and to The Wall Street Journal for giving me the opportunity to contribute to such an incredible publication. But more than anything, I hope this serves as a reminder that you never know who is watching. If you’ve been thinking about starting that blog, launching that social media page, starting a YouTube channel, or simply sharing something you’re passionate about, go for it. You never know where it might lead. You never know who might see your work. And you never know when an opportunity might come from it. Keep creating. Keep putting yourself out there. You never know who is watching! Here is a link to the article: https://lnkd.in/gys8_HZ8
-
Tript Singh Lamba liked thisTript Singh Lamba liked thisLinkedIn Fam... Well, the cat's out of the bag (or rather, the rocket has left the launch pad). 🚀 Last week I officially started my dream job as a Principal Technical Recruiter here at SpaceX — though my automated LinkedIn profile update beat me to the announcement. Lol. Better late than never — I'm beyond excited to be joining an incredible team on a mission I genuinely believe in: to make life multiplanetary. I'm also diving into a brand new industry, which has me equal parts energized and humbled. But really, how hard could it be? It's not like it's rocket science... oh, wait. 🤓 I'm so grateful for everyone who reached out last week with "likes" and kind words of encouragement, before I'd even said anything myself. That's just the kind of network family I'm blessed to have in all of you. 💙
-
Tript Singh Lamba liked thisTript Singh Lamba liked thisNew piece: The Harness, the Horse, or the Hay. The number of applications a person uses to do their job is collapsing to roughly one. If that's right, there are only three places to stand. The Horse is the model. Four or five companies on the planet are the horse, and if you're asking whether you should become one, you are not one. The Harness is the application layer for one rider and the entire job that rider does. There is no partial harness. You do all the jobs or you're a feature somebody ships in a sprint once their customers ask. The Hay is everything that keeps the other two running. Enormous market, completely different playbook. Almost nobody picks the wrong category. What kills people is picking the right one and running the other one's playbook. Also in here: why the model makers won't take the application layer, what Claudeforce actually tells us, and where the labor play and the services wave fit.
-
Tript Singh Lamba liked thisI closed with 𝗠𝗶𝗹𝗲𝘆 𝗖𝘆𝗿𝘂𝘀. On purpose. Stanford University Graduate School of Business had me teach a session in their CPO program this month. Thank you to the faculty and to the room for the hours. The arc I gave them: I have never thought about my career in titles. I have thought about it in hills. Capability growth, not title progression. A hill is a career chapter that forces you to learn something new. It has four parts, and I made the class run all four. -𝗧𝗲𝗿𝗿𝗮𝗶𝗻. 𝗪𝗵𝗮𝘁 𝘄𝗼𝗿𝗹𝗱 𝘄𝗲𝗿𝗲 𝘆𝗼𝘂 𝗶𝗻? The operating model you were actually inside, not the org chart you drew. -𝗧𝗮𝗿𝗴𝗲𝘁. 𝗪𝗵𝗮𝘁 𝘄𝗲𝗿𝗲 𝘆𝗼𝘂 𝘁𝗿𝘆𝗶𝗻𝗴 𝘁𝗼 𝗮𝗰𝗰𝗼𝗺𝗽𝗹𝗶𝘀𝗵? The OKR, or the BHAG that looked unreasonable the day you wrote it. -𝗧𝘄𝗶𝘀𝘁. 𝗪𝗵𝗮𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝗱 𝘁𝗵𝗮𝘁 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝗱 𝘄𝗵𝗮𝘁 𝘆𝗼𝘂 𝗸𝗻𝗲𝘄? Usually the moment the thing you were best at stopped working, the failure or setback. -𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆. 𝗪𝗵𝗮𝘁 𝗱𝗶𝗱 𝗶𝘁 𝘁𝗲𝗮𝗰𝗵 𝘆𝗼𝘂? Stated humbly, because the real ones are usually a little embarrassing. Then the part I believe most, which is not in the framework. The summit is never visible from the bottom. That is the feature, not the flaw. Not knowing whether the top sits at 600 meters or 3,000 is the thing that gets me out of bed. Every hill I have picked has been unfamiliar terrain, with at least one slope on it I had no business being on. And when I do get to the top, I have about three breaths in me. Standing on a summit is the least interesting thing you can do on a mountain. Usually, the moment I can finally see the top, I am already looking at the next hill. 𝗖𝗵𝗼𝗼𝘀𝗲 𝘁𝗵𝗲 𝗵𝗶𝗹𝗹 𝗳𝗼𝗿 𝘄𝗵𝗮𝘁 𝗶𝘁 𝘄𝗶𝗹𝗹 𝗳𝗼𝗿𝗰𝗲 𝘆𝗼𝘂 𝘁𝗼 𝗹𝗲𝗮𝗿𝗻. So the last slide was a 2009 clip of Miley Cyrus singing 𝗧𝗵𝗲 𝗖𝗹𝗶𝗺𝗯. I am aware of how that plays in a room of CPOs. I did it anyway. https://lnkd.in/gzkNhedF Because the whole song is one argument, and it is the same one I had just spent an hour making. It is not about how fast you get there. It is not about what is waiting on the other side. It is the climb. Sixteen years, and we are still writing frameworks to say what she said in a chorus. So here is the question I left them with, and the one I will leave here. 𝙒𝙝𝙖𝙩 𝙞𝙨 𝙮𝙤𝙪𝙧 𝙣𝙚𝙭𝙩 𝙝𝙞𝙡𝙡? 𝓐𝓷𝓭 𝓽𝓱𝓮 𝓱𝓪𝓻𝓭 𝓹𝓪𝓻𝓽: 𝓲𝓼 𝓲𝓽 𝓽𝓸𝓸 𝓼𝓲𝓶𝓲𝓵𝓪𝓻 𝓽𝓸 𝓽𝓱𝓮 𝓵𝓪𝓼𝓽 𝓸𝓷𝓮? 𝓘𝓼 𝓲𝓽 𝓳𝓾𝓼𝓽 𝓽𝓱𝓮 𝓼𝓪𝓶𝓮 𝓬𝓵𝓲𝓶𝓫 𝔀𝓲𝓽𝓱 𝓪 𝓫𝓲𝓰𝓰𝓮𝓻 𝓽𝓲𝓽𝓵𝓮 𝓸𝓷 𝓲𝓽?
Experience & Education
Licenses & Certifications
Volunteer Experience
-
Business Development Director
Ministry of Health, Kenya and Microclinc Technologies
- 3 months
Health
Selected as part of Skills4Afrika to lead a joint initiative between Kenyan ministry of Health and award winning EMR company Microclinic. I helped define their business model and deployment strategy for their flagship product Zidi in Kenya and other African countries. This award winning product has a variety of life saving measures for maternal and child health in rural Africa.
bit.ly/29GRLW1 -
Chair, Humanitarian Fund
Sikh Centre of Seattle
- 7 years
Social Services
Chair for Humanitarian Fund in Seattle based non-profit for education, healthcare and technology assistance
bit.ly/29zxZbx
Honors & Awards
-
Microsoft - Multiple Gold Star Awards, Management Excellence, Individual Contribution Awards, Teamwork and Collaboration awards
Microsoft
-
Microsoft HiPo (High Potential) program, Senior Leadership Bench
Microsoft - Windows and Device Group
This is a top 1% program for leadership succession and grooming where I now receive Senior Leader Bench coaching for execs and enjoy the benefits of a HiPo network
-
Microsoft representative - Industry working groups, external speaking/panel engagement in Mobile Congress, WBA etc
Microsoft and Industry Bodies
Languages
-
English
Native or bilingual proficiency
-
Punjabi
Native or bilingual proficiency
-
Hindi
Native or bilingual proficiency
Organizations
-
AngelList.co
Investor
-
IEEE
-
-
TIE
-
Recommendations received
14 people have recommended Tript Singh
Join now to viewOther similar profiles
Explore more posts
-
Stéphane Koch
❗(...) One former Google engineer, Abhishek Iyer, created dummy web pages that appeared only in Google’s index. When he later asked ChatGPT about them, the bot responded with information pulled from those very pages, suggesting that Google search snippets are indeed being repurposed in ChatGPT responses (like Meta, Apple and Perplexity). 🟥 [Comment] The work of Abhishek Iyer, and the examples provided by Backlinko, represent a gateway to exploiting "LLM Grooming" (flooding the internet with SEO-optimized content, ensuring its inclusion in the training datasets and search indexes used by artificial intelligence). To this end, by creating content structured to be indexed by Google's search engine, this technique can be used to alter the knowledge of generative artificial intelligence (data poisoning) with the aim of generating misinformation in its responses to specific questions, as was the case in the context of the Russian Pravda operation, identified by VIGINUM (more than three million articles intended to seed the knowledge of the main AGIs on the market had been produced), which raises a critical point about the vulnerabilities of AI systems to the manipulation of their information sources. ▶️ Abhishek Iyer: How does ChatGPT Search the Web: Data-Driven Investigation https://lnkd.in/ePbk7gH6 ▶️ Backlinko: ChatGPT Is Using Google Search – We Tested It https://lnkd.in/ePbk7gH6 ▶️ Exposing Pravda: How pro-Kremlin forces are poisoning AI models and rewriting Wikipedia https://lnkd.in/e-yNrGyX
-
Aditya Koushik V
What happens when you hire someone bcoz the industry is doing it? MIT Sloan research recently showed that companies who increased high-skill hiring in AI roles saw direct uplift in productivity and output per employee. But, when you blindly follow the industry or academia footprints, you are doomed - You raise money from investors / use company budget - Conduct interviews, On-board your team - When you do that, then you don't know what to get done when the person onboards - Then you artificially "Create a work" so that you keep the person occupied for 8 hours - You invest (Waste) 2 hours of your time in training that person and 2 hours in reviewing their work - Then you have to justify their productivity to your management / investors. So, you invest another 2 hours in making a presentation - You keep doing this for a month until you realize your work is not getting done - Then your actual productivity goes down and you get questioned by the management - Then you fire that person to keep investors happy - You achieved 1 month of productivity loss for you, the HR, the employee you recruited (you also made his life more miserable by hiring and firing him) and wasted others time in reviewing your actions. This is what happens when you do something bcoz it's an industry norm or tradition I'm not saying you shouldn't hire. I'm saying hire should be based on internal pain and should not be stimulated from external noise. When you are working day and night every day for weeks, then you know that it's the right time to delegate some portion of your work to be sustainable in the long term What others do, how the industry has been doing it for 50 years, what the Fortune 500 company does, what your neighbour does, what your competitor does - is NONE of your business Differentiating a Signal from Noise is an underrated skill in today's world #Foundertruths #MITsloan #AIproductivity #Teams #Scaleintelligently #Efficientgrowth #RealValue #Signal #Noise #Foundersmindset #Leadershiptruths #Startupfounders #B2B #SaaS #EnterpriseAI #Productivity #Businessclarity #Distraction
-
Abrar Siddiqui
Everyone is building AI agents. Very few are talking about what breaks in production. I joined Kuldeep Singh on Techie Talks with Kuldeep to discuss the real engineering challenges behind building SarahAI: Voice Latency, WhatsApp workflows, Non-deterministic AI behavior, token costs, and what it takes to make AI useful for SME founders. The big question we explored: How do you turn AI from an impressive demo into a reliable daily assistant? Full episode link is in the comments. #AI #AgenticAI #Startups #SoftwareEngineering #SarahAI
3 Comments -
Brian Laung Aoaeh, CFA
I have been speaking about "A Golden Age of Supply Chain Platforms, Networks, and Ecosystems" for about a decade. The key building blocks are now in place, and we are about to see an explosion of innovation where bits and atoms intersect. This is precisely the moment for which we have been building REFASHIOND Ventures: The Industrial Transformation Fund. Krenar's post below will give you an idea where things are heading. Happy to count him as an early investor.
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content