This Week in IT: OpenAI's New Model Family, AI Coding Tools, and a Supply Chain Attack You Should Know About
Welcome back to the itpick weekly news roundup — your no-fuss summary of the tech stories that actually matter for home users and small businesses. Here’s what caught our eye this week.
OpenAI Launches a New Family of AI Models — Three Flavours, One Big Step Up
OpenAI has announced GPT-5.6, a new family of AI models that comes in three versions designed for very different jobs. The top-tier model, Sol, is built for heavy lifting like complex coding and security research. Terra sits in the middle, aimed at business tasks such as document analysis and customer support tools. Luna is the lightweight option — quick, cheap, and well suited to everyday jobs like summarising content or drafting emails.
Right now, access is restricted to a small group of preview partners, so most of us won’t be able to try these immediately. But the direction is clear: OpenAI is moving towards offering different models for different budgets and use cases, much like choosing between a sports car and a reliable hatchback depending on what you need that day.
What this means for you: If you’re using AI tools like ChatGPT for work or personal productivity, expect more choice and (eventually) lower costs as these tiered models roll out more widely — tools like Claude are already doing something similar, so competition is heating up nicely.
AI Is Making Small Dev Teams Punch Well Above Their Weight
Here’s an interesting shift happening in tech companies right now. Anthropic, the company behind Claude, reportedly told its growth team to hire more product managers rather than extra engineers — because their AI coding assistant had effectively tripled the output of the engineers they already had. The bottleneck is no longer writing code; it’s deciding what to build.
This pattern is starting to show up in small businesses too. A developer or technically minded founder using an AI coding tool today can ship software at a pace that would have needed a full team just a couple of years ago. It’s a genuine productivity leap, not just hype.
What this means for you: If you run a small business and have been putting off building that internal tool or simple web app because you couldn’t afford a developer, now might be the time to revisit that. AI coding assistants are making it far more accessible — and if you’re already using productivity tools like Notion to manage your projects, pairing that with an AI coding tool could be a genuinely powerful combination.
Hijacked Code Packages Are Quietly Installing Infostealers on Your Devices
Security researchers have uncovered a nasty supply chain attack targeting developers. Cybercriminals hijacked legitimate packages used in popular development environments and quietly bundled in a Python-based infostealer — malware designed to harvest passwords, session tokens, and other sensitive data. The attack works across Windows, Linux, and macOS, and was cleverly designed to dodge some of the security checks built into newer versions of common developer tools.
Even if you’re not a developer yourself, this matters. A freelancer, IT contractor, or even a small business owner who uses open-source tools could be unknowingly running compromised software. It’s a reminder that threats don’t always arrive via dodgy email attachments.
What this means for you: Keep your development tools and dependencies updated, and if you’re storing passwords in a browser or a spreadsheet, it’s well worth switching to a dedicated password manager like NordPass — an infostealer’s first target is usually your saved credentials.
AI Agents Have a Growing Security Problem Worth Watching
Researchers and security professionals are raising louder alarms about “prompt injection” — a type of attack that manipulates AI agents into doing things their owners never intended. As more businesses plug AI into their workflows, the ways attackers can exploit these systems are multiplying. It’s still mostly an enterprise concern for now, but the underlying risk applies wherever AI agents are making decisions on your behalf.
What this means for you: If you’re using AI tools that can browse the web, send emails, or access your files autonomously, stay curious about what permissions you’re granting them — and keep an eye on this space as it develops quickly.
That’s your week in IT. Between OpenAI’s new model tiering, AI tools reshaping how small teams work, and some genuinely sneaky malware doing the rounds, it’s been a busy one. Stay safe out there, keep things updated, and we’ll be back next week with more of the same.