OpenAI has launched Codex, a new AI agent designed to execute specific work tasks rather than just chat. Unlike ChatGPT, which focuses on brainstorming, Codex handles multi-file projects, integrates with various tools, and manages repeatable workflows. It can build digital assistants, create dashboards, fix broken processes, and automate mundane duties. OpenAI states that Codex requires human oversight and clear direction, acting as an eager assistant rather than a replacement for human workers.
Professionals are finding ten specific use cases for Codex to boost productivity. Applications include creating daily briefs by aggregating calendar and email data, drafting weekly summaries and slide decks, and synthesizing research into decision memos. The tool also handles data management by cleaning files, standardizing fields, and building dashboards. Additionally, Codex supports financial reviews and product launches by pulling together campaign assets and identifying areas needing further review.
In the infrastructure space, Taiwan-based INFINITIX established a South Korea subsidiary to expand its presence in the Asian AI market. The company will showcase its AI-Stack platform at AI EXPO KOREA 2026 in Seoul, which uses proprietary CTAs technology to increase GPU utilization from about 30% to over 90%. INFINITIX also introduced ixCSP, an AI Cloud Service Platform that transforms infrastructure into a revenue-generating asset through models like GaaS, MaaS, and TaaS.
BT has partnered with Nscale and Rackspace to advance its UK sovereign AI ambitions by expanding its Sovereign Platform product suite. The group is building underlying AI data centre capacity through deals with Nscale, NVIDIA, and Rackspace Technologies. This initiative includes themes such as AI data centres, data sovereignty, and hybrid cloud services. The partnerships aim to support research commercialization and sovereign cloud capabilities across the United Kingdom.
Amplitude released AI Assistant, an embedded support agent that helps users inside digital products without sending them to separate channels. The tool uses user behavior data to guide tasks with step-by-step walkthroughs or complete workflows on their behalf. It monitors for signs of struggle and intervenes with contextual help before users disengage. The assistant connects to existing Amplitude tools like Guides, Surveys, and Session Replay to provide full case files if human support is needed.
Security concerns are rising as AI security leaders meet in Washington to discuss securing AI systems. The gathering featured experts from organizations like SANS, NIST, and OWASP to address how quickly vulnerabilities can be discovered in systems like Anthropic's Mythos. Concerns include zero-day vulnerabilities in traditional software and AI models, as well as the need for practical guidance on identifying AI-specific threats. The field remains fragmented with overlapping frameworks and little agreement on where to start securing AI systems.
Garry Tan, CEO of Y Combinator, discussed building software with AI agents during a Startup School session. He introduced the G stack, a tool designed to create an AI engineering team capable of handling tasks from product scoping to deployment. Tan demonstrated how AI agents can automate browser tasks like downloading tax documents and explained his Conductor Pattern for running multiple parallel code sessions. He emphasized the importance of review and iteration to ensure AI output aligns with product goals.
Akto secured AI agents through major platform partnerships with LangChain, Portkey, TrueFoundry, Arcade, and LiteLLM to expand AI agent security. These integrations embed runtime guardrails, continuous discovery, and observability directly into the infrastructure developers already use. The collaborations protect against prompt injection, sensitive data exposure, and policy violations across various AI agent gateways and runtimes. Akto ensures security travels with the agent rather than being bolted on afterward, making agentic runtime protection the default for engineering teams.
In the financial sector, the mortgage industry is exploring AI tools and new webcasts covering capital markets topics. Freddie Mac announced it is accepting VantageScore 4.0 for lenders to use. JazzX AI digital assistants are highlighted for coordinating complex decisions across processing, underwriting, and servicing to lower costs and improve loan quality. CI&T is offering specialty AI projects to help lenders integrate technology stacks with over 100 vendors or build custom AI tools for sales and underwriting.
Meanwhile, Australia's largest bank is cutting 120 jobs during its AI push and operational changes. Some 43 of the roles will be from Bankwest, CBA's subsidiary based in Western Australia, and include mobile lending managers. The Finance Sector Union stated that six of the cut roles were impacted by automation. This reduction is part of broader changes affecting the bank's workforce structure.
Finally, a debate continues on federal government AI regulation. Proponents argue that AI poses significant risks like job displacement and privacy invasion, requiring oversight to prevent social inequalities and unintended consequences. Opponents worry that overzealous rules could stifle innovation and give other countries a competitive edge. Recent incidents involving AI in political campaigns and deepfakes have raised concerns about malicious use, making the balance between innovation and regulation crucial.
Key Takeaways
- OpenAI launched Codex, an AI agent designed to execute specific work tasks like building dashboards and automating workflows, distinct from ChatGPT's conversational focus.
- INFINITIX opened a South Korea subsidiary to showcase its AI-Stack platform, which uses proprietary CTAs technology to boost GPU utilization from 30% to over 90%.
- BT partnered with Nscale, NVIDIA, and Rackspace to expand its UK sovereign AI ambitions and build underlying AI data centre capacity.
- Amplitude released AI Assistant, an embedded support agent that uses user behavior data to provide contextual help within digital products.
- AI security leaders met in Washington to address rising risks, including zero-day vulnerabilities in systems like Anthropic's Mythos and the need for unified frameworks.
- Garry Tan of Y Combinator introduced the G stack and Conductor Pattern to help teams build software using AI agents for tasks like automating browser interactions.
- Akto secured AI agents through partnerships with LangChain, Portkey, and others to embed runtime guardrails and protect against prompt injection and data exposure.
- The mortgage industry is adopting AI tools, with Freddie Mac accepting VantageScore 4.0 and JazzX assistants helping coordinate underwriting and servicing decisions.
- Australia's largest bank plans to cut 120 jobs as part of its AI push, with some roles in Bankwest directly impacted by automation.
- Global debate continues on federal government AI regulation, balancing concerns over job displacement and privacy against the risk of stifling innovation.
OpenAI launches Codex for automated work tasks
OpenAI introduced Codex, a new AI agent designed to execute specific work tasks instead of just chatting. Unlike ChatGPT, which helps with brainstorming, Codex handles multi-file projects, integrates with various tools, and manages repeatable workflows. It can build digital assistants, create dashboards, fix broken processes, and automate mundane duties. OpenAI states that Codex requires human oversight and clear direction, acting as an eager assistant rather than a replacement for human workers.
Ten ways professionals use Codex to boost productivity
OpenAI outlined ten specific use cases for its Codex model to help professionals automate tasks and generate content. Applications include creating a daily brief by aggregating calendar and email data, drafting weekly summaries and slide decks, and synthesizing research into decision memos. The tool also handles data management by cleaning files, standardizing fields, and building dashboards. Additionally, Codex supports financial reviews and product launches by pulling together campaign assets and identifying areas needing further review.
INFINITIX opens Korea subsidiary and shows AI infrastructure
Taiwan-based INFINITIX established a South Korea subsidiary to expand its presence in the Asian AI market. The company will showcase its AI-Stack platform at AI EXPO KOREA 2026 in Seoul, which uses proprietary CTAs technology to increase GPU utilization from about 30% to over 90%. INFINITIX also introduced ixCSP, an AI Cloud Service Platform that transforms infrastructure into a revenue-generating asset through models like GaaS, MaaS, and TaaS. Strategic partnerships with local distributors like Miruware and Tera Tec will help drive broader adoption.
INFINITIX expands AI infrastructure reach in Korea
INFINITIX launched a South Korea subsidiary to strengthen local capabilities and expand its footprint in the region's dynamic AI market. The company will display its unified AI-Stack platform at AI EXPO KOREA 2026, held from May 6 to 8 at COEX in Seoul. This platform addresses fragmented GPU resources by enabling centralized control across NVIDIA, AMD GPUs, and NPUs. INFINITIX formed partnerships with local distributors including Miruware, Tera Tec, DS &G, and NUMP to accelerate market entry and scale its AI infrastructure ecosystem.
Mortgage industry explores AI tools and new webcasts
The mortgage industry is discussing lender responsibilities, AI tools, and new webcasts covering capital markets topics. Freddie Mac announced it is accepting VantageScore 4.0 for lenders to use. JazzX AI digital assistants are highlighted for coordinating complex decisions across processing, underwriting, and servicing to lower costs and improve loan quality. CI&T is offering specialty AI projects to help lenders integrate technology stacks with over 100 vendors or build custom AI tools for sales and underwriting.
Amplitude releases AI Assistant for in-product support
Amplitude launched AI Assistant, an embedded support agent that helps users inside digital products without sending them to separate channels. The tool uses user behavior data to guide tasks with step-by-step walkthroughs or complete workflows on their behalf. It monitors for signs of struggle and intervenes with contextual help before users disengage. The assistant connects to existing Amplitude tools like Guides, Surveys, and Session Replay to provide full case files if human support is needed.
BT partners with Nscale and Rackspace for UK AI goals
BT commissioned Nscale and Rackspace to advance its UK sovereign AI ambitions by expanding its Sovereign Platform product suite. The group is building underlying AI data centre capacity through deals with Nscale, NVIDIA, and Rackspace Technologies. This initiative includes themes such as AI data centres, data sovereignty, and hybrid cloud services. The partnerships aim to support research commercialization and sovereign cloud capabilities across the United Kingdom.
Akto secures AI agents through major platform partnerships
Akto announced partnerships with LangChain, Portkey, TrueFoundry, Arcade, and LiteLLM to expand AI agent security. These integrations embed runtime guardrails, continuous discovery, and observability directly into the infrastructure developers already use. The collaborations protect against prompt injection, sensitive data exposure, and policy violations across various AI agent gateways and runtimes. Akto ensures security travels with the agent rather than being bolted on afterward, making agentic runtime protection the default for engineering teams.
Garry Tan discusses building software with AI agents
Garry Tan, CEO of Y Combinator, shared insights on building software with AI agents during a Startup School session. He introduced the G stack, a tool designed to create an AI engineering team capable of handling tasks from product scoping to deployment. Tan demonstrated how AI agents can automate browser tasks like downloading tax documents and explained his Conductor Pattern for running multiple parallel code sessions. He emphasized the importance of review and iteration to ensure AI output aligns with product goals.
AI security leaders meet in Washington amid rising risks
AI security leaders gathered in Washington to discuss securing AI systems as risks mount and new threats emerge. The meeting featured experts from organizations like SANS, NIST, and OWASP to address how quickly vulnerabilities can be discovered in systems like Anthropic's Mythos. Concerns include zero-day vulnerabilities in traditional software and AI models, as well as the need for practical guidance on identifying AI-specific threats. The field remains fragmented with overlapping frameworks and little agreement on where to start securing AI systems.
Australia's largest bank cuts 120 jobs during AI push
Australia's biggest bank plans to cut 120 more jobs as part of its AI push and operational changes. Some 43 of the roles will be from Bankwest, CBA's subsidiary based in Western Australia, and include mobile lending managers. The Finance Sector Union stated that six of the cut roles were impacted by automation. This reduction is part of broader changes affecting the bank's workforce structure.
Debate continues on federal government AI regulation
The rapid advancement of artificial intelligence has sparked a global debate on whether the federal government should regulate this powerful technology. Proponents argue that AI poses significant risks like job displacement and privacy invasion, requiring oversight to prevent social inequalities and unintended consequences. Opponents worry that overzealous rules could stifle innovation and give other countries a competitive edge. Recent incidents involving AI in political campaigns and deepfakes have raised concerns about malicious use, making the balance between innovation and regulation crucial.
Sources
- OpenAI Unveils Codex: Your AI Work Assistant
- Codex: 10 Ways to Work Smarter
- INFINITIX Launches Korea Subsidiary, Unveils AI Infra at AI EXPO KOREA 2026
- INFINITIX Launches Korea Subsidiary, Unveils AI Infra at AI EXPO KOREA 2026
- Specialty AI, Pooling, Correspondent Products; Lender Responsibility Opinion; Webcast Shows Incoming
- Amplitude launches AI Assistant for in-product support
- BT commissions Nscale, Rackspace to advance UK ‘sovereign’ AI ...
- Akto Announces Partnerships with Langchain, Portkey, TrueFoundry, Arcade, and LiteLLM to expand AI Agent Security
- Garry Tan on Building with AI Agents
- AI security leaders gather in Washington as risks mount—and Mythos raises the stakes
- Australia’s Biggest Bank to Cut 120 More Jobs Amid AI Push
- Should the federal government regulate artificial intelligence?
Comments
Please log in to post a comment.