Wickey

@Alfred_Lin The best who see your potential improvement points gives you suggestions at private, shows you examples, shares personal experiences, discusses the reasons of why to understand you and helps to understand the importance of the changes. :)

Is AI making it easier to start a company or harder to survive? This week, Professor Isley from Stanford's Technology Ventures Program (STVP) joins us in Innovator Cafe Episode 39 to unpack the opportunities and traps AI creates for founders. 📍 Why do Stanford students dare to start companies? Two words: permission and proximity. 📍 The most underrated #AI #risk isn't deepfakes, it's #trust. A Nature study Professor Isley co-authored found consumers punish brands whose ads land on misinformation sites, while most executives have no idea it's happening. 📍 Solo founder or team? A multi-year experiment tracking tens of thousands of founders found the "gritty, never-pivot" group pitched better short-term, but the flexible pivoters won long-term. Counterintuitive, and worth hearing. 📄 The papers behind the conversation: ① Nature: on how the ad-driven attention economy bankrolls misinformation, and why AI only makes it cheaper to fake. ② Persistence vs. Pivot: a multi-year A/B study (with a co-author at Wharton) tracking outcomes of tens of thousands of founders. Plus: the most overlooked resource in entrepreneurship: hint, it's not money. Dropping this week. Stay tuned.

I met @hao_sheng_ six months ago in an event. From our conversation, I noted that he was a founding engineer from Cresta as a builder. Even through he is an engineer founder never did the sales, his AI EA assistant startup @expertise_ai made to #1 in this category in Hubspot and reached 1 M ARR before the seed round. I was wondering the secret of growth until I saw this video below. If a shy engineer who is willing to sale his product without thinking his "regular" image, he is real all in what is his doing. Later on, I noted that he flied overnight to get an important customer by himself. That is why I started to think that he is a good founder to be invested. Here's the problem it's built for. according to Salesforce's own State of Sales data, Account Executives spend 72% of their week on non-selling work, and only 28% actually selling. but Expertise Assist isn't just trying to win back those hours. You can think that Expertise AI is an agentic platform built for B2B sales teams, think of it as "Cursor for Sales." Its flagship product, Expertise Assist, is an AI agent designed for Account Executives, seamlessly embedded into their sales workflow to help them close more deals, not just automate administrative tasks. Rather than simply saving time, it focuses on improving win rates by helping reps prioritize opportunities, manage pipelines, and execute sales activities more effectively. The company is operated by Chatsimple Inc., headquartered in Toronto. #AI #GTM #Innovation #Investment

The Biggest Security Risk Isn't the Attacker. It's Coordination - What the 2026 FIFA World Cup Teaches us See full article below for more details.

The people building the future of AI aren't just publishing papers. They're in this room. OpenAI. Anthropic. Microsoft. Stanford. Recursive. Greptile. Meet the researchers, founders, and operators shaping what's next at AGI Summit 2026. 📍 San Francisco 📅 July 18–19Tickets ↓Which speaker are you most excited to hear from?

👀QTS abandoning Digital Gateway QTS abandoning Digital Gateway does not signal weakening demand for #AI data centers. Rather, it shows that the data center industry is entering a new phase: shifting from capital-driven expansion to resource- and permitting-driven expansion. The future bottleneck for AI infrastructure will increasingly be less about GPUs and more about power, land, regulation, and community acceptance.

🌿Delfos Energy: The Virtual Engineer That Turned Asset Data into Action Most people think the energy transition is about building more renewable capacity. Delfos #Energy is betting on something different: that running existing assets well will matter as much as building new ones，and that what actually bottlenecks this industry was never the hardware. It's the expertise. Here's how the story unfolds. 1. renewable operators connect data from SCADA systems, sensors, weather feeds, maintenance logs, and other operational systems into the Delfos platform 2. instead of simply displaying dashboards, Delfos continuously analyzes asset performance using machine learning to detect abnormal behavior, early-stage failures, and hidden revenue losses 3. the platform doesn't just generate alerts, it identifies the likely root cause, prioritizes issues by financial impact, and recommends the next engineering action 4. engineers can interact with the platform in natural language, turning complex operational data into actionable decisions without manually analyzing thousands of signals 5. over time, Delfos aims to evolve from predictive analytics into a true AI virtual engineer that automates engineering workflows, maintenance planning, and operational decision-making across wind, solar, hydro, and battery storage assets. The problem behind all this is very real. The renewable workforce is shrinking while portfolios get harder to manage. Experienced engineers are retiring faster than they can be replaced, and wind, solar, and now battery storage keep piling more complexity onto operators who are already stretched thin. SCADA systems can tell you what happened. They can't tell you what it's costing you, or what to do next. Delfos turns that gap into a product, distilling the judgment of a seasoned performance engineer and making it available, continuously, across a thousand sites at once. Most companies in this space are still competing to build a nicer-looking monitoring dashboard. What Delfos is trying to replace is the engineer's mental model itself, pairing a proprietary machine-learning core for real-time performance analysis with a second AI layer, built on open-source LLMs plus domain-specific guardrails, aimed at automating the workflows around maintenance planning and reporting. Delfos's founders are a near-textbook founder-market fit: CTO Samuel Lima holds a Mechanical Engineering master's focused on wind energy and co-founded Energo Engenharia (2014-2018), one of Brazil's first wind turbine service providers, before bringing aviation-style predictive maintenance into wind energy. CEO Guilherme Studart holds four international-management master's degrees and cut his teeth in finance/procurement at logistics giant América Latina Logística and Volkswagen, giving the company its commercial and fundraising engine. Founded together in Barcelona in 2017, this pairing of field engineering expertise and international business execution has scaled Delfos to 15+ GW monitored, 1,000+ sites across 10+ countries, and €10M raised to date. This one is not planted but founder-market fit resonates with me because it reflects a broader enterprise challenge. In #cybersecurity and #compliance, we face the same problem: organizations have no shortage of data, dashboards, or alerts, they have a shortage of experienced practitioners who know how to interpret them and make the right decisions. Delfos isn't just building #AI for renewable energy; it's demonstrating how AI can capture domain expertise and make it available at scale. I believe we'll see the same pattern emerge across security, governance, audit, and many other knowledge-intensive industries.

Safety evaluation platform for robotics Recently, I discussed the evaluation in our podcast and feel that it is an interesting topic. This one may solve the problem so I’d like to share with my friends. SafeWorld provides an evaluation platform that helps robotics teams validate system performance, identify risks, and continuously improve their models through iterative evaluation with focus on robots operating around people. SafeWorld uses #AI and simulation to generate rare but critical scenarios that are difficult to capture through conventional testing. The platform continuously evaluates robotic systems under these scenarios, uncovers failure modes, quantifies #risk, and enables teams to iteratively improve system safety and reliability before deployment. Team-market fit: The founding team combines experienced entrepreneurs with leading robotics safety researchers. Co-founder Ding Zhao is a professor at Carnegie Mellon University and recipient of the NSF CAREER Award, with extensive research in robotics safety and trustworthy autonomy. Co-founders Kyle Wong and Simo Rachidi bring startup and AI product experience, combining deep technical expertise with product execution.

Anthropic didn't just open-source a #security tool. They open-sourced an opinionated blueprint for how AI agents can secure code autonomously. Threat modeling → Vulnerability discovery → Triage → Reporting → Patching. That's more than an AppSec workflow. It's a glimpse of how security engineering itself may evolve. The interesting question isn't whether #AI can find vulnerabilities anymore. It's how we build the orchestration, guardrails, and approval layers that allow AI to remediate them safely in production.

@manthanguptaa Are you in the US now?:)

Four trends that stood out to me at this year's AI Engineer World Fair • Memory, context engineering, context graphs, retrieval, and knowledge system companies were everywhere. It feels like everyone has realized that context, not the model, is the bottleneck. • Local AI is no longer a niche. The demand for ownership, privacy, lower latency, and model portability is becoming much stronger. • The vibe coding hype is cooling off. More talks focused on review loops, evals, testing, and human oversight than on replacing engineers. • "Software factories" is the new buzzword. Instead of asking AI to write code end-to-end, the focus is shifting toward building systems that continuously review, test, evaluate, debug, and improve software throughout its lifecycle.

Happy 250th Birthday, America. Thirty-two years ago, the Internet was beginning to reshape how we communicated, learned, and built businesses. Today, we're standing at the beginning of another transformation, the AI era. Over the next decade, AI will transform how we work, innovate, and solve problems. But the best technology doesn't replace human connection, it helps us appreciate it even more. Yet today reminds me that some of the most meaningful moments in life have nothing to do with technology. They're the neighborhood barbecues where everyone brings something to share. Children running around with sparklers. Friends laughing over a picnic table. Families gathering before the fireworks begin. Neighbors who may not know each other's last names, but still wave and wish one another a "Happy Fourth." As someone who spends much of my time thinking about #AI, #Cybersecurity, and the future of technology, I hope we continue building technologies that strengthen human connection rather than replace it. The Internet connected us. AI will amplify what we're capable of. But community, kindness, gratitude, and the simple joy of spending time together will always remain wonderfully human. Happy Independence Day! 🇺🇸❤️