In partnership with

Login or Subscribe to participate

On 26 October, Proton told selected staff they'll be re-interviewed on how they use AI, for the jobs they already hold.

The same day, Microsoft offered more than a quarter of its workforce money to leave. Every Microsoft employee opening their inbox that morning:

Read the two emails together and you get a pretty clear picture of what "good at AI" is going to mean at work. Both are below, right after a short intermission.

A word from two tools that won't re-interview you.

Presented by TigerData

Store It. Search It. Analyze It. One Postgres

Your AI app generates three kinds of data at once: events, vector embeddings, and the analytics you run on top of them. The default move is three systems, three sets of tooling, and pipelines to keep them in sync.

TimescaleDB handles all three in the Postgres you already run.

Hypertables ingest events at scale. pgvector and pgvectorscale store and search embeddings. Continuous aggregates keep analytics live without re-querying everything.

Same SQL, same tools, one system to operate. No pipelines to sync, no drift, no second database to feed.

Point it at your heaviest workload and watch it stay fast as data grows. It's still Postgres. Start on Tiger Cloud and get $1000 in credits.

Presented by Wispr Flow

Turn today's calls into tomorrow's proposal

Wispr Flow Notetaker brings your meetings into Claude or ChatGPT through a built-in connector. Ask for a proposal draft or a deal recap, and your AI works from an accurate record of what was said.

And we're back. Proton first.

Proton wants receipts

Proton emailed selected employees saying they'll be re-interviewed about their AI use next month. 404 Media got hold of the email, as Tom's Guide reports.

My favourite line from the email, via TechRadar: "A list of tools alone will not demonstrate this."

Reality check: Proton says on Reddit this isn't about layoffs, that it's short on talent, still hiring, and wants to spot who needs extra support.

Fair enough, but the word "consequential" is doing a lot of heavy lifting in that email.

Microsoft and the "voluntary" exit offer

Big tech has a fancy name for everything, and goodbyes are no exception.

Microsoft's CPO Amy Coleman announced a Voluntary Exit Offer. Around 60,000 people are eligible, which is more than a quarter of the company.

That covers people at level 70 and below in teams that opted in, which in plain words means almost everyone short of top leadership, including some VPs (but not corporate VPs), across the US, Canada, the UK and India.

In India the deal is one month's base salary for every completed year of service, plus health insurance and career transition support, and you have until December 1 to decide.

So six years at Microsoft gets you six months of base pay and a slightly awkward farewell cake.

The official line: Microsoft says AI isn't driving this, performance plays no part, and it expects only a small share of people to accept.

I'd believe that a lot more if the leaked list of participating teams looked different.

According to an internal memo seen by the Seattle Times, Xbox, Windows, Copilot, HR and finance are fully in, AI, legal and cloud are partly in, and marketing, sales and security sat it out.

Put them together: both emails are asking one question in two different ways.

What has AI done for your work, and could your team run with fewer people?

So before your own company sends its version, let's get you prepared. But first, a quick stop.

Tools on my radar

A developer built his wife a job search app after her search dragged on for almost 10 months.

Dossier finds postings, writes a tailored resume and cover letter for each one and tracks every application on a board, from draft to offer.

She landed a job, so he open-sourced it, and it's free.

Walk a room or a whole building with your iPhone for a few minutes and it turns the scan into a photorealistic 3D world, true to size.

The app is free to download, scans run on credits, and it's iPhone only for now.

Private AI that runs on your own phone and computer, so your chats never leave your devices.

Handy for anything you'd rather not paste into a cloud chatbot, like salary sheets or client contracts.

The re-interview survival kit

1. Your receipts

Scroll back a month in your ChatGPT or Claude history. Find three times AI changed how you did your job.

The last line is where an interviewer will push, so make it the most specific one on the card.

3. One flop and a no

Proton asks what didn't work, so keep one story where AI got it wrong and you caught it, like a made-up revenue figure you spotted right before hitting send.

Its Reddit post also says using AI well means knowing when to keep it out. Pick one task you deliberately didn't hand over, salary data for example, and be ready to explain why.

4. The cousin test

Read your cards to the cousin who still thinks you "do computers." If he's lost after a minute, rewrite.

5. Mock re-interview

❝

Paste this into Claude:

Interview me about how I use AI at work. You're from another department, so you don't know my team's jargon. Ask one question at a time about what I changed, the results, what failed and when I chose not to use AI. Push back on vague answers. After ten questions, score me 1 to 10 and tell me the weakest answer and where I can improve.

6. The Friday log

Ten minutes every Friday, one line on task, tool, time saved, what you checked. By March you'll have an appraisal-ready list.

The bottom line

Proton's email asks a fair question, and most of us would struggle to answer it on the spot.

If you can't fill three proof cards from the last month, it's much better to find that out on a Friday with Claude than in a room with HR.

Hit reply with your best proof card, and tell me how you'd answer if your manager asked you tomorrow.

Until next time,
Vaibhav 🤝🏻

Reply

Avatar

or to participate