Claude gets quicker, ChatGPT gets helpers that never sleep, and a two-minute safety check: This Week, Plainly #9

Hello again. It’s been a busy week in AI: both big companies launched something new within a day of each other. Here’s what happened, in plain English, and what (if anything) you need to do.

1. Claude just got quicker, including on the free plan

What happened: On 28 September Anthropic, the company behind Claude, released Claude Sonnet 5.5. It’s now the standard version of Claude for free and Pro users. Anthropic says it replies more than 30% faster and is better at everyday jobs like writing documents, making slides and working with spreadsheets.

Should you care? ⚪ No action needed. Claude switches over by itself. You might notice answers arriving a little faster. As we said about Opus 5.5 last week, the names and numbers are mostly for techies. The skills you’re learning work the same on every version.

2. OpenAI’s big developer day: AI helpers that keep working

What happened: On 29 September OpenAI, the company behind ChatGPT, held its yearly developer conference, called DevDay. The headline was “dots”: AI helpers that live inside ChatGPT and work on a goal for you around the clock, using their own web browser. OpenAI also launched a new model, GPT-6.1 Sol, which is available across ChatGPT.

Should you care? ⚪ Not yet, for most people. Dots are only for ChatGPT’s most expensive plans ($200 a month and up) and business accounts. An AI that acts for you while you’re not watching needs a lot of trust, so we’d wait and see how it goes for the early users. The new model arrives by itself, so there’s nothing to do.

Worth knowing: If you’re thinking about a top-tier ChatGPT Pro plan, check what it includes before you pay. OpenAI changed what new subscribers get this week.

3. 🔐 A new way to check your ChatGPT account is safe

What happened: ChatGPT now has a security history. You can see recent sign-ins and sign-outs, plus any changes to your security settings.

Should you care? 🟢 Yes. Take two minutes to do this today. Go to Settings → Security and look for your sign-in history. If you see a device or location you don’t recognise, change your password and turn on multi-factor authentication (a code sent to your phone as well as your password). If it’s already on: well done.

What happened: Which? is warning about a scam where criminals phone out of the blue pretending to be your bank, the police or the Financial Conduct Authority. They say your account is at risk, then ask you to move to a WhatsApp or FaceTime video call and share your screen. Once you do, they can see the security codes your bank texts you, and use them to take your money.

Should you care? 🔴 Yes, this is the one to remember. No real bank or police officer will ever ask you to share your screen or install an app so they can “help”.

If you get a call like this, hang up
Call your bank back on the number on the back of your card, ideally from a different phone
Never read out or share a security code that’s been sent to you

Report it at reportfraud.police.uk or call 0300 123 2040, and tell your bank straight away.

5. The big picture: AI is helping to build the next AI

What happened: Anthropic says Claude is now doing a large share of the research and engineering work on future versions of Claude, with people overseeing it.

Should you care? ⚪ Nothing to do, but it helps explain why new versions keep arriving so quickly. Two new Claude versions in two weeks is a good example. It’s also why we focus on skills that last (asking clearly, checking answers, staying safe) rather than the latest button or feature.

💡 Tip of the week

Review what your AI remembers about you. Both ChatGPT (Settings → Personalization → Memory) and Claude (Settings, look for Memory) let you see and delete what they’ve remembered. It takes a minute and it’s often quite interesting.

Forward this to someone who’s been meaning to “get into AI”. They can sign up free at deadeasyai.com/news.