AI Exponential Thinker

I've interviewed around 150 candidates for ML/AI Engineer positions across various levels over the last 12 months. You guys asked me to share my takeaways, so here they are: I've noticed patterns in what separates good engineers from truly great ones. 1. A focus on understanding and evaluating business value The best ML engineers don't just write code - they solve problems with a clear purpose. I've seen too many candidates who lacked the ability to connect their technical decisions to the business outcomes they were meant to serve. Great engineers can assess the value their work brings and articulate why their approach is the right one. 2. Achieving the optimal balance between traditional and modern ML/AI approaches Some candidates rely heavily on "fancy" AI techniques, over-engineering solutions when simpler methods would suffice. Others stick to outdated tools and miss the chance to deliver quick, impactful proofs of concept. The most impressive candidates demonstrated both depth and flexibility - knowing when to use cutting-edge methods and when a classic ML algorithm was the better choice for the problem at hand. 3. Turning stakeholders' ideas into actionable roadmaps A skill I rarely saw, but deeply value, is the ability to take a vague stakeholder idea and transform it into a clear technical roadmap. Many candidates struggled to scope realistic PoCs or prioritize work to show early value. Great engineers know how to break down big-picture goals into actionable steps and communicate a plan that aligns with both technical and business priorities. 4. Clear and impactful communication of technical concepts Too often, candidates failed to explain their thought processes in a way that made sense to stakeholders or collaborators. The best engineers I've interviewed could simplify complex concepts, tailor their language to their audience and make their ideas resonate - whether with a fellow engineer or a non-technical executive. A great AI engineers care about the impact of their solutions and design the most effective approach to deliver real value, not just complete tasks.

You only need to read four books to truly get what’s going on in ML and data engineering: - Fundamentals of Data Engineering by Joe Reis - Designing Data Intensive Applications by Martin Kleppmann - AI engineering by Chip Huyen - Designing Machine Learning Systems by Chip Huyen If you read these four technical books and then read these four books on leadership and soft skills, you’ll be well on your way to massive success! - Radical Candor - Atomic Habits - How to Win Friends and Influence People - The Body Keeps Score What books would you recommend?

Listening isn’t waiting for your turn to speak. Listening is the ability to make others feel heard. One of the ways to do that is to practice using these three statements when someone is telling you something: “Tell me more." "Go on.” “What else?” If you try this, you will quickly learn how much effort it takes to pay attention to what the person you’re talking to is saying and resist the urge to interrupt, correct, or fix what they say. And that’s why it’s a skill. Because all skills take effort. And in this case, the effort is really worth it. Learn more about the skills we teach in The Optimism Library: https://lnkd.in/ewVTWbis

Sam Altman in 2023: "It's totally hopeless to compete with us" ?? When asked by an Indian investor how could an Indian team build something substantial with $10 million. Fast forward to 2025, DeepSeek spent $5.6 million, outcompete OpenAI, and did this as a side project. In the race for innovation, it's not the size of the wallet, but the sharpness of the mind that wins. P.S. for more great stuff, check out ??linas.substack.com??, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

Dans l’univers #Technologie & #Business, ou j’opère depuis deux décennies à travers des géographies diverses et des cultures variées, l’outil diplomatique est fort présent dans les négociations, les coopérations, la gestion des relations mais aussi l’implantation de nouvelles activités voir la dissolution de partenariats… Dans quelques heures, la date 20 Janvier 2025 marquera a new page de #USA & #Monde En ce jour ou nous fêtons 5 ans de AI Exponential Thinker et avec un nombre important #expert #Technology #diplomat #Busines #investment #Education, nous ouvrons le chapitre “Tech Diplomacy Society” Fière d’être #Tunisienne et citoyenne du #Monde dans un moment ou l’histoire de l’#Humanité va devoir revoir l’ensemble de ses acquis et modèles, j’ouvre une page et j’invite celles et ceux qui ont la #conviction #conscience et #confiance d’un meilleur avenir ou #Technologies doivent servir #Humanité et pas #injustice & l’effacement de l’autre en le réduisant à une valeur marchande de #données et #image #Diplomatie #geopolitoque #humanité #justicesociale #paix #securité #cooperation #Multilaterale #Ambassadeurs #internationales Mon article dessous sur #Tech #diplomat & Diplomatie Moderne #IA #Numerique #souveraineté

The *real* impact of Gen AI on human learning: a summary of the findings & implications of the five most recent & important research papers. More and more of us are using generic AI tools like ChatGPT, Claude, Gemini & Co-Pilot as learning tools, but what impact are they having on how humans learn? In this week's blog post I cut through the noise by providing an overview of the findings and implications of the five most recent & important research papers. Here's the TLDR: ?? Study 1: How ChatGPT impacts student engagement from a systematic review and meta-analysis study (2024) > AI improves cognitive & emotional engagement, especially in struggling students. However, it limits the development of higher order skills compared with traditional learning methods. ?? Study 2: The Impact of Large Language Models on Programming Education and Student Learning Outcomes (2024) > AI enhances the speed of task performance but hinders foundational skill development, especially in beginners. ?? Study 3: Let GPT be a Math Tutor: Teaching Math Word Problem Solvers with Customized Exercise Generation (2023) > Customised AI exercises can improve problem-solving and skill retention, but "over supportive" generic AI tools negatively impact long term retention and higher-order skill development. ?? Study 4: AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking (2024) > Cognitive offloading to AI reduces learners' critical thinking engagement (r = -0.75) compared with traditional methods. ?? Study 5: Beware of Metacognitive Laziness: Effects of Generative Artificial Intelligence on Motivation, Self-Regulated Learning, and Performance (2024) > AI tools boosts learning confidence but negatively impacts actual learning outcomes. Practical Implications: →Limit the Use of Generic AI in Foundational Learning: During the early stages of learning, focus on manual problem-solving and hands-on practice to build strong conceptual knowledge and foundational skills before incorporating AI tools. → Treat Generic AI as a Learning Activity, Not a Teacher: Use tools like ChatGPT, Claude, Co-Pilot & Gemini for practice, problem sets and exploration. Encourage learners to engage with AI-generated outputs by solving problems, analysing solutions, and critiquing its responses to develop deeper understanding and critical thinking skills. → Advocate for Pedagogy-First AI Models & Tools: Be wary of the risks of using generic AI for learning. Seek out "pedagogy-first" AI systems that are designed by both engineers and educators, with evidence-based learning principles at their core. For more info, including a vision of what pedagogy-first AI systems might look like, check out the full blog post. Happy experimenting! Phil ?? #AI #AIed #education #chatgpt #claude #gemini #copilot #edtech #openAI https://lnkd.in/esMCyp8e

Final letter from the President's Council of Advisors on Science and Technology (PCAST): "Future Opportunities for Science and Technology to Impact the Nation" News story on the final session of PCAST: https://lnkd.in/gd5nZ7p5 White House Office of Science and Technology Policy Frances Arnold Maria Zuber Arati Prabhakar

The only way to shape the #Future of #prosperity #peace #inclusity #stability is by “Make #AI your ALLY” thank you to HEC Paris for hosting an amazing event with leaders from exceptional organizations #Waymo #OpenAI #MIT #Harvard and inviting me as speaker to such #audience of #Researchers #leaders, Thank you ???? We need to shape a new wave of Modern #Diplomacy with #Tech #Diplomat for #bilateral #multilateral #cooperation #geopolitics #economics #societies #Citizen are worried in these times of frequent #Crisis and we have a responsibility to stand up ?? and contribute to build this near #Future of #AGI #AIagents #GenAI #Autonomousvehicle #AR #VR #Mixedreality #biotech #Technologies AI Exponential Thinker

Dont invest in AI Agents without reading this Anthropic report Here's a comprehensive breakdown of the research... Often most automation that we use daily can be easily done with a few LLMs and APIs and that's something Anthropic too is trying to convey This research shares what they think are AI agents and why you should not always build one. ?? We recently covered Google's whitepaper where we broke down how Google's vision of AI Agents. Now, after reading both Google's and Anthropic's take - It is safe to say that Google's paper was more on What AI Agents. Whereas Anthorpic's take is much more on Why and When should you should use AI Agents. Here's a brief breakdown from their research: ?? Agents vs. Workflows: Agents are dynamic systems where LLMs direct their own processes and tool usage, while workflows follow predefined paths. Agents shine when flexibility and decision-making are key. ?? Core Parts of AI Agents: 1. Augmented LLMs 2. Tools used by the Agumented LLMs 3. Environment 4. Memory ?? Key Workflow for Agents: - Prompt Chaining: Breaking tasks into sequential steps for higher accuracy. - Routing: Directing inputs to specialized tasks for better performance. - Parallelization: Running tasks simultaneously for speed or diverse outputs. - Orchestrator-Workers: A central LLM delegating tasks to worker LLMs. - Evaluator-Optimizer: Iterative refinement by multiple processes for polished results. ?? When to Use Agents: You don't always need to use Agents, Often your automation can be easily done using a few automation workflow tools like N8N and other commercial tools. Here are a few problems where you should use AI Agents: - Open-ended problems require flexibility. - Tasks where decision-making scales with complexity. - Environments with trusted autonomy and clear feedback loops. ?? Few Frameworks given by Anthropic to Consider: - LangGraph (LangChain) - Amazon Bedrock's AI Agent Framework - Rivet and Vellum for GUI-based workflow building ?? Key Takeaway:? - Success isn’t about building the most complex system—it’s about building the right system.? - Start simple, measure performance, and add complexity only when it demonstrably improves outcomes. - Sometimes without understanding the core aspect of an agent we redundantly pile up too much code from a few frameworks which Often leads to redundant code being piled up without understanding. - Hence they specifically aimed this research to bring more clarity to people who are trying to build AI agents for their businesses. What are your thoughts about this report? Let me know your thoughts in the comments below ?? Please make sure to, ?? Share ?? React ?? Comment to help more people learn P.S. Let me know if you want a detailed summary of any other reports or papers

#exponential #technologies are impacting the #Science of #Diplomacy and we need #Techdiplomats #Ambassadors who will shape new #AI #policies, foster multilateral cooperations, promote #peace & #stability for #ALL in times of #geopolitical crisis & uncertainty AI Exponential Thinker