IBM Technology
IBM TechnologyVideo
- Format:
- YouTube channel
- New videos:
- Daily
- Typical length:
- about 15 minutes, not counting Shorts
- Latest video:
About this show
Whether it’s AI, automation, cybersecurity, data science, DevOps, quantum computing or anything in between, we provide educational content on the biggest topics in tech. Subscribe to build your skillset, learn about new trends, and gain insights from IBM experts.
New videos every Sunday through Friday!
--------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
This channel follows the IBM Social Computing Guidelines.
Videos come from the channel's public YouTube feed. The artwork is the newest video thumbnail. Last updated . Names, artwork and text belong to their publishers.
Cosmos
Similar by topic and creator
- The PrimeTime
Shares the topic Technology. - Just Have a Think
Shares the topic Technology. - Neso Academy
Shares the topic Technology.
Videos
Is AI Really Superintelligence? ANI, AGI & ASI Explained
· 13 min
Learn more about Artificial Superintelligence? here → https://ibm.biz/~ma2KWmdlF Is today’s AI truly superintelligent, or are we overstating its capabilities? Jeff Crume explains intelligence through perception, thinking, learning, and action, then compares ANI, AGI, and ASI. Learn why current AI remains narrow and why greater capability demands security, governance, testing, and human control. 00:00 – Ask If AI Is Really Superintelligence 00:29 – Define What Intelligence Actually Means 02:23 – Understand Perception in Intelligent Systems 03:31 – Explore AI Thinking and Reasoning 04:28 – See How Intelligent Systems Learn 05:29 – Turn Intelligence into Action 06:41 – Define Artificial Intelligence 07:27 – Compare ANI, AGI & ASI 10:21 – See Why Today’s AI Is Not Superintelligent 11:01 – Control AI Capability with Safety and Governance AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~zn1XTTn5G AI was used in the creation of the transcript and metadata for this video. #ai #agi #aialignment #superintelligence --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetSoftware Supply Chain Security: What AI Coding Changes
· 9 min
Learn more about Supply Chain Security here → https://ibm.biz/~k518Rj9sC Your code is only part of what reaches production. Cedric Clyburn explains how dependencies, CI/CD pipelines, open source, and AI-generated code expand the software supply chain. Learn how SBOMs, automated checks, independent review, and provenance reduce risk. 00:00 – See What Really Enters Your Application 01:00 – Define the Modern Software Supply Chain 02:02 – Understand Dependency Risk with Log4j 03:03 – Protect Package Registries from Ecosystem Poisoning 04:00 – See How Build Pipelines Become Attack Vectors 04:47 – Treat AI Coding Agents as Software Suppliers 06:01 – Shift Software Supply Chain Security Left 06:31 – Build an SBOM for Every Application 07:04 – Automate Security Checks and Verify AI Code 08:08 – Prove Software Provenance Before Production AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~rqEKFmJiG AI was used in the creation of the transcript and metadata for this video. #supplychainsecurity #softwaresecurity #aicodingagent #aicoding --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetAI Model Debugging: Find the First Divergence Across AI Accelerators
· 14 min
Learn more about AI Debugging here → https://ibm.biz/~i26jCJsKx The same AI model can produce different results across accelerators. Kaoutar El Maghraoui explains AI model debugging by finding the first observed divergence between executions. Learn how tensor comparisons, tolerances, and minimal tests help isolate failures. 00:00 – Debug Different AI Model Results Across Accelerators 00:53 – Trace Errors Through the AI Software Stack 02:40 – Separate Expected Variation from Correctness Bugs 03:05 – See How Small Differences Propagate Through Models 04:45 – Localize the Difference Before Diagnosing It 05:52 – Build a Validated Reference Execution 06:50 – Find the First Observed Divergence 08:42 – Compare Tensors with Numerical Tolerances 10:23 – Isolate the Operation with a Minimal Test 12:32 – Turn the Fix into a Regression Test Continue your education in the course Debugging Across the Dataflow Stack on Coursera here → https://ibm.biz/~NdKCGTusI AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~Tsdh7IC6m AI was used in the creation of the transcript and metadata for this video. #aimodel #aiaccelerator #aiengineering #machinelearning --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetSecurity's New Reality Check
· 41 sec
Everyone's a target. The question isn't if you'll be breached—it's how fast you can contain it. Learn the containment-first security mindset. --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetHuang's Nvidia Legacy
· 40 sec
Jensen Huang built Nvidia over 30 years with relentless passion. Can anyone replace him? Explore the challenge facing AI's most iconic leader. --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetJensen Huang exits NVIDIA & does Claude Fable 5.2 count as AGI?
· 35 min
Visit Mixture of Experts podcast page to get more AI content → https://ibm.biz/~MikQp78Aa Who can replace a founder like Jensen Huang? This week on Mixture of Experts, host David Zax is joined by Derek Robertson, Bri Kopecki and Kiyu Gabriel. First, NVIDIA CEO Jensen Huang resigns after more than three decades leading a company whose long-term investments helped shape the AI industry. The panel discusses whether the long-term bets Huang made, from CUDA to high-bandwidth networking, can keep NVIDIA ahead without him. Next, Anthropic releases Claude Fable 5.2, renewing questions about artificial general intelligence (AGI). A surprising demonstration of the model leads the panel into a debate about human creativity. If AI-generated art makes you feel something, does it matter who—or what—created it? Finally, a special one-on-one conversation examines Confluent's Streamhouse announcement and what access to real-time business data could mean for AI agents. All that and more on Mixture of Experts. 0:00 – Intro 01:02 – Jensen Huang resigns from NVIDIA 09:57 – Claude Fable 5.2 & what counts as AGI 28:16 – What is a Streamhouse & why it matters "The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity. AI tools may be used to transcribe this episode and support selected stages of the production process. All AI-assisted content is reviewed by the production team before publication." AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~WVEDzwD8N #nvidia #claudefable #agi AI was used in the creation of the transcript and metadata for this video.
Not rated yetWhy AI Agents Need RAG + Observability
· 9 min
Learn more about Why Observability Is Essential for AI Agents here → https://ibm.biz/~MBafoTJOL Your AI agent knows the system, but what about now? Matt Aiken explains how RAG and observability give AI agents documented knowledge plus live metrics, logs, and traces. Learn how better operational context helps teams investigate issues with evidence. 00:00 – See What AI Agents Still Don't Know 00:37 – Give AI Agents Two Types of Context 00:46 – Ground AI Agents with RAG 01:47 – Add Live Context with Observability 03:06 – Investigate a Hybrid Transaction with AI 04:12 – Add Logs, Metrics & Traces to the Investigation 05:23 – Connect RAG Knowledge to Live Evidence 06:22 – Turn AI Context into Evidence, Impact & Next Steps 07:35 – Keep Human Operators in Control 08:09 – Combine RAG + Observability for Better AI Agents AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~9KyFFw5rY AI was used in the creation of the transcript and metadata for this video. #aiagents #retrievalaugmentedgeneration #observability #devops --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetThe $6B Binance heist, Meta’s AI bug tests and looming quantum deadlines
· 32 min
Visit Security Intelligence podcast page to get more AI content → https://ibm.biz/~93M8GNpAP Hackers spent months planning a $6 billion cryptocurrency heist. How did they get past security controls, and what can defenders learn from the attack? This week on Security Intelligence, Seth Glasgow, Dimple Ahluwalia and Suja Viswesan join host Patrick Austin to unpack three major cybersecurity stories. First, the Binance heist reveals how attackers exploited employees and internal systems to steal billions in cryptocurrency. The panel examines the sophisticated tactics behind the attack and what security teams can learn. Next, Meta is using AI to catch software bugs before they reach production. As AI takes on more of the testing process, how much human oversight is enough? Finally, IBM introduces Quantum Safe to help organizations prepare for post-quantum cryptography. With attackers potentially harvesting encrypted data today to decrypt later, the panel discusses the challenges of identifying risks and prioritizing migration. All that and more on Security Intelligence. 00:00 - Intro 01:05 - Inside the $6B Binance crypto heist 12:56 - Meta's AI tests catch bugs early 22:25 - IBM introduces Quantum Safe suite "The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity. AI tools may be used to transcribe this episode and support selected stages of the production process. All AI-assisted content is reviewed by the production team before publication." AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~T4vJnnCuG #socialengineering #aisecurity #incidentresponse AI was used in the creation of the transcript and metadata for this video.
Not rated yetWhat Is Model Distillation? How Smaller AI Models Learn
· 13 min
Learn more about Knowledge Distillation here → https://ibm.biz/~9OAE6ZHmW Small AI models can learn from much larger ones. Aadam Mulla explains model distillation, where student models learn from a teacher’s soft labels and dark knowledge. Learn how distillation balances AI capability, speed, cost, and deployment. 00:00 – Understand What Model Distillation Does 00:36 – See Why Smaller AI Models Matter 01:15 – Connect Model Parameters to Memory and Speed 02:24 – Meet the Teacher and Student AI Models 02:50 – See How LLMs Predict the Next Token 04:59 – Train the Teacher Model with Hard Labels 07:26 – Train the Student with Soft Labels 09:12 – Understand Why Soft Labels Teach More 10:04 – Explore Dark Knowledge in Model Distillation 12:09 – Balance AI Capability with Model Efficiency AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~xkl5jbfCz AI was used in the creation of the transcript and metadata for this video. #aimodels #llm #machinelearning #aiengineering --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetSpec-Driven Development vs. Waterfall: What's Different for AI Coding?
· 7 min
Learn more about Spec-Driven Development here → https://ibm.biz/~zVL5K4NRM AI coding is making software specifications relevant again. Cedric Clyburn compares Spec-Driven Development with Waterfall and explains how evolving specs give AI agents better context. Learn why clearer specs can mean fewer prompts, faster iterations, and better code 00:00 – See Why AI Coding Needs Better Specs 00:43 – Define Spec-Driven Development 01:06 – Understand the Waterfall Development Model 01:38 – See Why Waterfall Specs Go Stale 02:47 – Ask If Spec-Driven Development Is Just Waterfall 03:09 – Build with an Iterative Spec-Driven Loop 03:39 – Refine an AI Coding Spec as You Build 04:18 – Keep Specifications Version-Controlled with Code 05:05 – Give AI Agents Better Context & Less Ambiguity 06:36 – Add Security, Compliance & Governance to Specs AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~lsxtAZPcw AI was used in the creation of the transcript and metadata for this video. #specdrivendevelopment #aicoding #aiagents #waterfall #softwareengineering --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetWhy AI Models Run Faster on Some Chips: The KV Cache Effect
· 10 min
Learn more about KV Cache here → https://ibm.biz/~6LPSw4Y8N The same AI model can run differently across chips. Kaoutar El Maghraoui explains how GPUs, TPUs, and AI accelerators store and move KV cache during LLM inference. Learn why memory architecture and data movement can shape AI performance. 00:00 – Understand Why LLMs Need a KV Cache 00:46 – See How KV Cache Reuses Keys and Values 01:39 – Balance KV Cache Capacity and Fast Memory 02:41 – Follow KV Cache Through GPU Memory 04:41 – Compare TPU Memory and Compiler Orchestration 06:44 – Explore Dataflow Accelerator Memory 08:34 – Compare How AI Accelerators Move KV Cache 09:05 – Connect KV Cache Movement to LLM Performance Continue your education in the course Memory Systems in Dataflow Architectures on Coursera here → https://ibm.biz/~I7KAjOCAb AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~modQneKWV AI was used in the creation of the transcript and metadata for this video. #kvcache #llm #aimodels #aiaccelerator --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetAI fixes AI in MIT study, Mistral's open model & Google Genie 4
· 35 min
Visit Mixture of Experts podcast page to get more AI content → https://ibm.biz/~Rz77ywquX How close are we to AI that can improve itself? This week on Mixture of Experts, guest host Aili McConnon is joined by Chris Hay, Sascha Brodsky and Kiyu Gabriel tackle that question and more from the latest AI headlines. First, MIT researchers demonstrate a technique that uses AI to modify another model's weights. The panel considers how close this brings us to recursive self-improvement and whether reward hacking poses a more immediate risk. Next, Mistral releases a trillion-parameter open-weight model while NVIDIA expands its open model offerings. With open models approaching frontier-level capabilities, will developers still need to rely on proprietary AI? Finally, Google DeepMind introduces Genie 4, its new world model. The panel debates how much these models actually understand about physics and what happens when a convincing simulation gets the real world wrong. All that and more on Mixture of Experts. 00:00 – Intro 01:08 - AI self-improvement and MIT's research 16:28 - New open models from Mistral and NVIDIA 25:27 - Google’s Genie 4 simulates 3D worlds "The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity. AI tools may be used to transcribe this episode and support selected stages of the production process. All AI-assisted content is reviewed by the production team before publication." AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~AUQOseCFm #aiselfimprovement #openweightsAImodels #AIworldmodels AI was used in the creation of the transcript and metadata for this video.
Not rated yetAre Vector Databases Still Needed? RAG, AI Agents & Search
· 10 min
Learn more about Vector Databases here → https://ibm.biz/~aHY62ka9p Do AI applications still need vector databases? Martin Keen explores RAG, AI agents, embeddings, semantic search, and context windows to explain when vector databases matter and when simpler retrieval approaches work just as well. Learn how to choose the right retrieval architecture for your workload. 00:00 - Are Vector Databases Still Needed in the Age of AI Agents? 00:29 - What Vector Databases Actually Do: Embeddings & Semantic Search 01:15 - The Retrieval Pyramid: Choosing the Right Search Strategy 01:49 - 3 Reasons You Might Not Need a Vector Database 02:01 - Agentic Search vs Vector Search: Can Keyword Search Win? 03:08 - Do Long Context Windows Make Retrieval Obsolete? 03:33 - The Hidden Costs of Embeddings and Vector Indexes 05:27 - When Vector Search Outperforms Keyword Search 07:22 - Vector Index vs Vector Database: What's the Difference? 08:18 - How to Decide the Right Retrieval Approach for Your AI System AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~YCTMzkcZz AI was used in the creation of the transcript and metadata for this video. #vectordatabases #retrievalaugmentedgeneration #aiagents #vectorsearch #ai --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetDid the Knicks’ phishing test go too far? Plus, hackers pose as IT in Teams
· 30 min
Visit Security Intelligence podcast page to get more AI content → https://ibm.biz/~vrweMGEy6 Click to claim your free Knicks championship ring, and you might end up in security training. First, MSG sends staff an offer to pick their own championship ring, and everyone who clicks lands in mandatory security training. Timely lures catch real behavior, but the panel debates whether a test this real makes employees wary of the emails they actually need. Then, Arctic Wolf traces more than 400 incidents of attackers flooding inboxes with sign-up emails, then posing as IT support in Teams chats. As one panelist points out, real IT already has access to your machine, so it should never need you to open a remote session. Finally, CISA confirms attackers are exploiting two critical PHP flaws, and 263,000 end-of-life instances in the US alone will never get a patch. Do you know which versions you’re running? All that and more on Security Intelligence. 0:00 – Intro 1:04 – MSG's championship ring phishing test 8:34 – Hackers pose as IT in Microsoft Teams 19:21 – PHP2Shell flaws exploited in the wild "The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity. AI tools may be used to transcribe this episode and support selected stages of the production process. All AI-assisted content is reviewed by the production team before publication." AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~9IfKz4JuD #phishingsimulation #cybersecurity #vulnerabilitymanagement AI was used in the creation of the transcript and metadata for this video.
Not rated yetCan Your AI Use Sensitive Data Safely? 5 Capabilities You Need
· 11 min
Learn more about Digital Sovereignty here → https://ibm.biz/~Wxx0y9d9w Sensitive data makes AI powerful, but also risky. Jeff Crume explains five essential capabilities for using sensitive data safely across AI systems, agents, tools, and RAG pipelines. Learn how visibility, classification, governance, access controls, and evidence help secure AI-driven workflows. 00:00 - Why AI Requires Machine‑Speed Data Controls 00:18 - The Sensitive Data Conundrum: Useful but Risky 01:02 - Five Foundational Controls for Safe AI Adoption 01:13 - Visibility: You Can’t Protect What You Can’t See 02:16 - Automated Classification and Sensitivity Tagging 03:59 - Data Residency, Sovereignty, and Localization Explained 05:46 - Jurisdictional Complexity in Cloud Environments 07:12 - Access Control and the Four As of IAM 08:53 - Audit Trails and Continuous Evidence for Regulators 10:27 - Trustworthy AI Requires Trustworthy Data 11:06 - Turning Sensitive Data Into a Strategic AI Advantage AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~OuRghwRQo AI was used in the creation of the transcript and metadata for this video. #aisecurity #digitalsovereignty #aidata #cybersecurity --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetContext-Aware Coding: Give AI Agents the Right Context
· 6 min
Learn more about Agentic Coding here → https://ibm.biz/~YeRLqinv2 A good prompt isn't enough for good AI code. Legare Kerrison explains how context helps AI coding agents understand your repository, dependencies, tests, and intent. Learn how to give agents the right context without overwhelming them. 00:00 – Give AI Coding Agents Better Context 00:37 – See Why One Prompt Isn't Enough 01:23 – Give Agents the Right Code & Repository Structure 02:17 – Ground AI Coding with Dependency Versions 02:43 – Use Tests as an Agent Feedback Loop 03:00 – Add Documentation & Developer Intent 03:42 – Let AI Agents Ask for Missing Context 04:02 – Interview, Explore & Plan Before Coding 04:58 – Avoid Overloading the AI Context Window 05:45 – Keep Agentic Coding Context Focused AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~G97V5h6dy AI was used in the creation of the transcript and metadata for this video. #aicodingagent #agenticcoding #aicoding #aiagents --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetWhat Is torch.compile? Dynamo, PyTorch & Faster AI Models Explained
· 6 min
Learn more about Faster AI Models here → https://ibm.biz/~9RJjTpaDF One line can change how your AI model runs. Avery Blanchard explains how torch.compile, Dynamo, AOT Autograd, and Inductor optimize PyTorch models. Learn how graph breaks, recompilation, and kernel fusion affect AI performance. 00:00 – See What torch.compile Does to PyTorch Models 00:20 – Understand the Three torch.compile Stages 01:09 – See How PyTorch Starts Lazy Compilation 01:44 – Capture AI Model Operations with Dynamo 02:42 – Prepare PyTorch Graphs with AOT Autograd 03:32 – Generate Optimized Code with Inductor 04:35 – Understand Graph Breaks in torch.compile 05:21 – Avoid Recompilation & Benchmark Correctly 05:53 – Connect Dynamo, AOT Autograd & Inductor 06:10 – Diagnose Why torch.compile Isn't Faster Continue your education in the course Compilation for Dataflow on Coursera here → ibm.biz/~E46ZY0koj AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~j9Ety2SXE AI was used in the creation of the transcript and metadata for this video. #pytorch #aimodels #machinelearning #aiengineering --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetWhen Security Patches Leak
· 58 sec
What happens when vulnerabilities leak before patches are ready? Discover why one point of failure puts 70% of the web at risk. --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetAI Exposes Legacy Code
· 35 sec
AI is forcing teams to revisit abandoned code. What dependencies are lurking? Time to modernize your stack before vulnerabilities strike. --------------------------------------------------------------------------------------------------------- Find us on YouTube: 🔵 IBM Technology: https://www.youtube.com/@IBMTechnology 🔵 IBM: https://www.youtube.com/@IBM 🔵 IBM Developer: https://youtube.com/@IBMDeveloperAdvocates
Not rated yetAI exposes a major web security flaw, plus a new kind of voice model
· 39 min
Visit Mixture of Experts podcast page to get more AI content → https://ibm.biz/~7OaIQxINM This week on Mixture of Experts, host David Zax is joined by Ash Minhas, Judith Aquino and Jessica Nahulan to discuss how AI is testing the systems we rely on, from cybersecurity to model evaluation. First, the panel examines PHP2Shell and what happens when AI can find vulnerabilities faster than organizations can fix them. Can our aging software infrastructure keep up? Next, Block's Dessa introduces a voice model trained entirely on speech that can listen and talk simultaneously. The panel discusses what makes this approach different and how to evaluate a model designed to converse more like a human. Then, a new cultural riddles benchmark reveals gaps in how language models interpret cultural context, prompting a broader debate about whether today's benchmarks reward guessing over admitting uncertainty. Plus, IBM Field CTO Kiyu Gabriel joins from IBM TechXchange 2026 to discuss Streamhouse and how AI agents are changing the demands on enterprise data infrastructure. All that and more on Mixture of Experts. 00:00 – Intro 01:18 – AI discovers major PHP2Shell flaws 14:30 – Dessa: An AI model trained on speech 22:38 – Cultural riddles reveal model blind spots 30:22 – IBM TechXchange: Data for the agent era "The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity. AI tools may be used to transcribe this episode and support selected stages of the production process. All AI-assisted content is reviewed by the production team before publication." AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~z5xHzLfpb #aicybersecurity #ibmtechexchange2026 #aiagents AI was used in the creation of the transcript and metadata for this video.
Not rated yet