Claude is way more powerful than most people realize. These 3 features prove it. The people getting real value are using features sitting right there in the interface that almost nobody touches. Kevin broke down the full workflow: > Turn a receipt photo into a live expense
AUTOMATION
-
Learning-Based Industrial Package Stacking Tackles a 3D Planning Problem
By
–
Fun acronym for an exciting approach to learning efficient industrial package stacking — a challenge for physical AI because it’s a 3D version of the popular Tetris game, known to be NP-Hard.
-
Shopify Agent Resolvas Automates Returns Within Defined Spending Limits
By
–
LOS TICKETS DE DEVOLUCIONES SE COMEN TU SOPORTE
— Nico (@nicos_ai) 26 août 2026
Resolvas es un agente de IA que vive dentro de Shopify:
→ resuelve el 60-80% de tickets de reembolso y cambio automáticamente
→ en menos de 15 segundos
→ tú defines el límite de gasto que puede aprobar
Plan gratis y sin… pic.twitter.com/Ae9ohkB2xPRETURN TICKETS STRAIN YOUR SUPPORT TEAM Resolvas is an AI agent that lives inside Shopify: → it automatically resolves 60–80% of refund and exchange tickets
→ in under 15 seconds
→ you set the spending limit it can approve [Traduit de FR vers EN] -
Reusable AI Systems Can Automate Repeated Campaign Tasks
By
–
Crazy to me how many people only use AI for single tasks. If you're about to do a task and you know you're going to do it more times in the next 2 months, you should instead build the system for that task. Ex: you want to launch a paid ad campaign. Instead of making the ads
-
Scheduled ChatGPT and Claude Workflows Update Repositories Through MCP
By
–
it’s still in progress but some stuff like granola is as simple as a scheduled workflow in ChatGPT/Claude that uses the MCP to look at recent stuff and updates the repo
-
Databricks Demo Builds Store Operations Agent With MCP and MLflow
By
–
[DEMO] Give an AI agent the right context, tools, and guardrails, and it can do a lot more than answer questions.
— Databricks (@databricks) 26 août 2026
Watch how to build a store operations agent that pulls from enterprise context and MCP servers, traces behavior in MLflow, turns follow-up work into a recurring… pic.twitter.com/AzKTdmuJkx[DEMO] Give an AI agent the right context, tools, and guardrails, and it can do a lot more than answer questions. Watch how to build a store operations agent that pulls from enterprise context and MCP servers, traces behavior in MLflow, turns follow-up work into a recurring
-
Specialized AI Agents Can Divide Tasks Across Multiple Roles
By
–
The truth is that having several specialized agents makes much more sense than asking one to do everything. [Traduit de ES vers EN]
-
Xiaomi Uses Autonomous Humanoid Robots for EV Assembly
By
–
Humanoid #Robots Join Xiaomi's #EV Factory for #Autonomous Assembly
— Ronald van Loon (@Ronald_vanLoon) 26 août 2026
by @IntEngineering#Robotics #Engineering #ArtificialIntelligence #Innovation #Technology pic.twitter.com/F1IPyUdg1VHumanoid #Robots Join Xiaomi's #EV Factory for #Autonomous Assembly
by @IntEngineering #Robotics #Engineering #ArtificialIntelligence #Innovation #Technology -
ChatGPT Work Plans a Disneyland Trip From One Prompt
By
–
🆕 ChatGPT quietly stopped being a chatbot.
— Futurepedia – Learn to Leverage AI (@futurepedia_io) 26 août 2026
Kevin puts ChatGPT's new "Work" mode through its paces; and it's less "answer my question," more "go do the whole job." A few moments that landed:
🔹 One prompt → a complete Disneyland trip: hotel comparison, day-by-day plan, budget,… pic.twitter.com/j2b8qxTBDJChatGPT quietly stopped being a chatbot. Kevin puts ChatGPT's new "Work" mode through its paces; and it's less "answer my question," more "go do the whole job." A few moments that landed: One prompt → a complete Disneyland trip: hotel comparison, day-by-day plan, budget,
-
TDK SensEI Agents Automate Maintenance Scheduling and Parts Ordering
By
–
Ordering parts, checking inventory, scheduling technicians. Maintenance managers do these manually every time a machine flags a problem. TDK SensEI's purpose-built agents handle them automatically. Join TDK SensEI at IMTS 2026 and find out more. Partner content with TDK SensEI.… pic.twitter.com/jYVpQRElZd
— Lucian Fogoros (@fogoros) 26 août 2026Ordering parts, checking inventory, scheduling technicians. Maintenance managers do these manually every time a machine flags a problem. TDK SensEI's purpose-built agents handle them automatically. Join TDK SensEI at IMTS 2026 and find out more. Partner content with TDK SensEI.