TT Aspiring

@3BodyProblem My FYP's all Optimus and Atlas right now 🤖 Thirty years of quiet cost curves and suddenly every VC's a hardware expert.

Wondering why your timeline is full of awesome vids featuring humanoid robots? The cost of the main components (lidar, vision systems, actuators) of robots has fallen by nearly 90% over the past 30 years.

@efipm I'd pay the ticket just for the RBI pivot story. Surviving a full regulatory shutdown and refocusing? Most European founders facing DORA and MiCA wouldn't know where to start.

Paytm Expands to Europe via Luxembourg Hired Nasir Zubairi to lead regional ops. Sharma at Nexus: AI fraud detection, credit scoring, payment infrastructure. India's fintech model = third template alongside US/China. Luxembourg as EU gateway for global fintechs buff.ly/2dM8in1

@SpirosMargaris This hits home for insurtech. We've spent years proving AI value to underwriters who only ask 'how many staff does it replace?' Real signal is pricing accuracy and faster claims triage. Start with the outcome.

Many companies are struggling to prove AI ROI because they are measuring the wrong things. The biggest impact of AI may not appear first in quarterly numbers, but in workflow redesign, decision speed and organizational capability. That is the key shift. AI is not just another software upgrade. It changes how companies operate. time.com/article/2026/0… @microsoft @time

@cleartechtoday @Veeam Proper interested in the non-human identity track 👍 Most insurers I know are handing out machine permissions like it's 2019. Not sustainable. How's your firm handling it?

📌 This week #VeeamOn2026 is happening in NYC You know how I love 💕 me a big ole security focus and @Veeam is bringing it The key security theme of VeeamON 2026 is "Data & AI Trust CONVERGE," focusing on the intersection of data resilience, identity protection, and securing autonomous systems in the Agentic AI era.  The core security sub-themes featured across the conference tracks include: 💕 Convergence of Resilience and AI Security:Breaking down silos between data backup and cybersecurity to manage real-world AI risks.  💕 Securing the "Agentic" AI Pipeline:Protecting the underlying Retrieval-Augmented Generation (RAG) pipelines, training data, and vector stores utilized by autonomous AI agents.  
 💕Identity Governance & Non-Human Threat Mitigation: Securing and recovering Active Directory, Entra ID, and machine identities as AI agents inherit broad enterprise permissions.  💕 AI-Driven Ransomware Response:Countering advanced cybercriminals who use AI to bypass detection and quietly extract data.  
 💕 Infrastructure Hardening & Immutability:Expert-level strategies for configuring secure backups, stress-testing immutability, and proving recovery readiness before an incident.  
 💕 Dark Data Visibility & Compliance: Using centralized data commands to classify unmapped, unmanaged enterprise data to maintain strict cloud sovereignty and compliance.  #AISecurity #CloudSecurity

@fintech_germany Yep. One bad AI decision with someone else's money and you're explaining yourself to the FCA. Human check isn't optional. I watched it happen.

The biggest AI mistake: Trying to automate everything instantly. Winners start with human-in-the-loop and scale trust. Blog: startuprad.io/post/ai-employ… Watch: youtu.be/bnrWAfiOfj8

@trudydarwin @Forbes Forbes has this right. Everyone wants the AI headline until something misfires. In insurtech we draft guardrails six months too late. What's your experience getting governance to stick?

Many overlook #AI guardrails says @forbes bit.ly/4uoPO5Q #AIGovernance #AIRisk #Tech #TechPR #CyberSecurity #EnterpriseAI #Security #DataProtection

@antgrasso Exactly this. We baked DIFC compliance into our AI models from day one. We hit UK and EU markets with zero rebuild. Standards-first sounds slower, but it isn't.

Global standards define market entry long before products compete, setting technical conditions that shape certification and interoperability. Companies that align engineering with compliance reduce redesign cycles and enter markets with fewer barriers. Microblog @antgrasso

@FGraillot McKinsey makes data mapping look lovely, but when your policy admin is held together by hope and middleware, agents can't save it. Saw the same mess in Dubai. Did anyone at the event actually have a budget for the foundation work?

Could Agentic AI unlock the modernization of insurance legacy systems? → linkedin.com/posts/florian-… McKinsey recently published an article detailing how Agentic AI could actually boost the modernization of insurance legacy systems.

@WindowOnTech @wef Everyone's an expert now. Show me your 2013 calls and I'll show you mine.

Rt @wef 5 ways to make the electric vehicle battery more sustainable wef.ch/3b3Wsbv #ClimateChange #ElectricVehicles

@ValerioCapraro So the brain respects phrase boundaries before guessing the next word? Makes sense. Probably why ChatGPT still can't read a balance sheet.

Important Nature Neuroscience paper shows how humans differ from LLMs. Many people currently believe that humans are just next-word predictors, like LLMs. But this new paper by Zou, Poeppel and Ding suggests something more interesting. The human brain does predict words. But it does not predict every word with the same precision. Prediction is constrained by linguistic structure. When a word continues the current phrase, brain activity tracks word surprisal in a way that resembles an LLM. But when a word crosses a major phrase boundary, the match weakens. In other words, the brain does not simply ask: “What is the next word?” It also asks: “What structure am I currently building?” This challenges one of the most common biases in today’s technological world: the belief that human language works like a large language model. The answer is: no. Human language is not just next-token prediction. * Paper in the first reply

@WorldRobotDay He is. Hinton's been saying this for years. AlphaZero parallel isn't perfect but the direction's spot on. In insurance we're already generating synthetic claims data to train on. It gets interesting when it stops imitating.

A voice to trust!

@Corix_JC @Forbes Same story in insurtech. Everyone wants the AI pitch. Nobody budgets for validation until something leaks. Seen this before? 😬

#AI Guardrails >> The New Frontier For Enterprise #Risk buff.ly/FfA1j4a #tech #data #digital #business #leadership #management #riskmanagement #CIO #CTO #CDO #CEO #CRO #digitalstrategy #AIstrategy #genAI #AgenticAI #digitaltransformation #disruption @Forbes @BetaMoroney @ChuckDBrooks @MikeFlache @Khulood_Almani @IngridVasiliu @JonBelsher @Timothy_Hughes @NafisAlam @YuHelenYu @RLDI_Lamy @MCLynd @OlKonol_oa @RigneySec @BenRothke @HWillert @AndrewinContact @TerenceLeungSF @Shi4Tech @enilev @Nicochan33 @SabineVdL @bimedotcom @IanLJones98 @FrRonconi @RiccardoBua @Eli_Krumova @NYIke @HaroldSinnott @SallyEaves @GlobalIQX @AntGrasso @AkwyZ @DinisGuarda @PVynckier @YvesMulkers @HaleChris @BillMew @RobMay70 @NigelTozer @m49D4ch3lly @Ronald_vanLoon @jeancayeux @NevilleGaunt @asokan_telecom @cybersecboardrm @RVP @QuePasaChico

@TheHighway2AI @EYnews @antgrasso Four pillars sounds ambitious. Most banks I know can barely keep their APIs upright. Which pillar gets ignored first, do you reckon?

Open banking will change the financial services landscape, but its global acceleration and adoption will be shaped by four critical pillars. @EYnews go.ey.com/2ACWPea rt @antgrasso #OpenBanking #Finserv #Fintech

@dhinchcliffe @OpenAI The shift from demo to deploy is where most AI projects die. I keep finding the same thing in insurtech. OpenAI just admitted the plumbing matters more than the model. Who saw that coming?

The biggest problem in enterprise AI is now getting into production. And doing so without detonating your architecture, governance, security, workflows, budgets, or employees. ⛓️ So @OpenAI launching a deployment company is a more revealing signal than most people realize. 📡 This is OpenAI basically saying: “Now it’s time to go big. Start wiring frontier AI directly throughout the enterprise.” And really, this tracks exactly with what I’m seeing in the CIO world as well. Most large enterprises already have access to plenty of strong models and AI infra. What they lack is: • AI workflow redesign chops • Operational experience • Integration capability • Governance • Trustworthy deployment patterns • Agent oversight • Cost controls • Organizational readiness That’s why still so many “AI transformations” still look like disconnected pilots and executive theater. The really strategic part? OpenAI is moving beyond being just a model company and trying to become part of the enterprise operating layer itself. 🔥 Because whoever helps deploy the AI: • Influences the stack • Shapes standards • Embeds into workflows • Controls the telemetry • Becomes extremely hard to remove later And now they’re doing it with investment firms, consultancies, and GSIs, which is basically the traditional enterprise power structure. That’s no accident. This is the market maturing in real time: from model competition → To deployment competition → Eventually to operational AI dominance. CIOs must pay attention. The next phase of AI will not be won by the firms with (only) the flashiest AI models. It’ll be won by the organizations that can safely speed-run operational AI, repeatedly, and at scale across the business while everyone else is still doing the basics.

@Corix_JC @ITPro @EmmaWoollacott @EYnews Seeing it daily in insurtech. Trust isn't a feature you ship later. It's the product. Boards are learning that the expensive way. Always the customer who pays.

#AI adoption is accelerating in the #UK, but ‘#trust is not keeping pace’ buff.ly/wdhjkhv @ITPro @EmmaWoollacott @EYNews #tech #data #digital #innovation #leadership #management #governance #strategy #CIO #CTO #CDO #CEO #digitalstrategy #AIstrategy #genAI #AgenticAI #digitaltransformation #disruption #digitaltrust

@PrivatEquityGuy I'd take a dusty brokerage with sticky commissions over a slick AI frontend any day. Gies built $5B from air filters because patience beats hype. What are we overlooking in insurtech?

Learning the art of bargain hunting from the late billionaire Michael Heisley, founder of Chicago’s Heico Companies. "Success leaves clues..." Gies started in the early 2000s and has since completed more than 80 acquisitions across over 20 countries, focusing entirely on the most "boring" corners of the industrial economy. Filtration, medical, safety, energy, industrial solutions, air, and Seakeeper... the last of which makes motion-control devices for marine applications.

@DivesTech Looks a good session. Did insurtech come up, or was it all NVIDIA multiples and no plumbing? I'd have asked about underwriting automation 👀

Great to be in London discussing tech stocks and the AI Revolution 🔥🎯🇬🇧💂🐂🍿🏆

@ingliguori Spot on. It’s not the model, it’s the orchestration. I care more about whether an agent can bind cover and chase data than whatever LLM sits underneath. Anyone actually running this in production?

Agentic AI = systems, not models 👇 • Intelligence (LLMs) • Memory • Tools • Feedback loops 🚀 That’s how AI moves from thinking → to doing Via Giuliano Liguori (@ingliguori) #AI #AIAgents #AgenticAI

@cleartechtoday The 'no retraining' bit is what matters. We haven't got budget to fine-tune models. Is this native to OpenShift AI or an add-on?

🌟🌟🌟Guardrails-as-a-service🌟🌟🌟 What is it? How does it work? Red Hat Summit 2026 will take place in Atlanta, Georgia, at the Georgia World Congress Center from May 11-14, 2026 Part of the security stack being discussed is Guardrails-as-a-Service, often referred to through the TrustyAI Guardrails framework, provides a modular, platform-level security layer for Large Language Models (LLMs). It allows organizations to enforce safety, compliance, and security policies on AI interactions without needing to retrain or modify the underlying models.  The system functions as an inline inspection layer between the user and the generative model. It typically includes:  🛤️ Guardrails Orchestrator: Acts as the central "traffic controller". It receives incoming requests, routes them through specific detectors, and only forwards them to the model if they are deemed 🛤️ Detector Services: Independent microservices that contain specific logic for identifying risks like PII (personally identifiable information), toxic language, or prompt injections.  🛤️ Guardrails Gateway: Provides standardized endpoints (like chat completion) and allows developers to preset specific detector pipelines for different applications.  #AISecurity Cc: @RedHat

@PJTheEconomist Corp tax at 25%, non-dom on the way out, NI up. Founders I know are quietly running the same numbers for Dubai. Hard to blame them, but what's the endgame here?

Britain isn’t the country you think it is thenewworld.co.uk/james-ball-bri… Lovely interview with James Ball. Thank you. His concluding paras sum it up...