
Hey folks,
My wifi has always been average. Video calls froze mid sentence. Occasional pages freeze. Someone would start an upload and everything ground to a halt.
Every time I opened the Google Home app, it told me my network was blazing fast and ran a speed test showing 903 Mbps.
I was getting between 50 and 120 Mbps at my desk.
After a few prompts with Claude Cowork driving the diagnosis of my wifi, and some walking around the apartment, I was getting 328 Mbps at my desk. If you want to fix your wifi, read below.
Let's dive in..

What I prompted
I want to show you the first message, because people assume there is a trick to this and there is not.

The only 2 things in there that mattered:
"From where I am sitting right now." My router's app tests the router. Nobody sits inside the router. The number that matters is the one in the chair you work.
"Do not suggest fixes until you have measured something." Ask an AI why your wifi is slow and it will tell you. It will give you a tidy, plausible list, you will pick one, change it, and feel vaguely better for a week. Ask it to measure first and you get your house instead of the internet's average house.
It came back with real numbers from my desk, and they did not match the app. That was the moment I realised I had been reading the wrong instrument for years and had never bothered to check it against reality.

Then Claude sent me for a walk
The second prompt was even simpler. I picked up the laptop, moved, and typed:
Me: I am now standing in the hall. Run exactly the same test again so we can compare. Then I will move somewhere else and we will do it again.
I did that in 4 places. Desk, hall, next to the router, far bedroom. Then back to the desk to check the first reading was not a fluke, because these numbers bounce around and one reading proves nothing.
What comes out the other end is a map. Not an opinion about your wifi, a map of it. And a map tells you things a dashboard never will.

My equipment was in an objectively terrible place. On the floor, in a corner, behind cardboard boxes and a plastic tub. I knew exactly where it was. I had walked past it every day for a year.
Adding coverage in the wrong place adds nothing. When I did eventually add a second unit, I first put it near the existing one, because that is where there was a spare socket. It did nothing at all. 2 devices sitting in the same spot are not coverage. Moved to the middle of the apartment, it lifted the far end substantially.
None of that is specialist knowledge. It is just that almost nobody measures, so almost nobody finds out.

Screenshot your router settings for Claude

This is the part I would tell a friend about first.
At some point the free stuff runs out and you have to open your router's settings page. Rows of dropdowns full of acronyms, numbers with no units, toggles that sound important, and no indication of which one is your problem. For most people this is where it ends, and fairly so.
I could read that page. What I could not do was hold every measurement from the last 2 days in my head and work out which of options to move first. So I stopped reading it and screenshotted it instead.
Me: Here are screenshots of my router's settings pages. Based on everything you measured around the apartment, tell me exactly which of these settings to change and what to change them to. Explain what each one actually does in plain English. And flag which changes are safe and which could knock the whole house offline, because I have to work today.
It read the page and gave me advice about the specific box in my hallway rather than routers in general, because it could see which options I actually had. Which dropdown. Which value. Why that one, given what it had measured at my desk 20 minutes earlier.
If you do not know what any of it means, say so in the prompt. Add "I do not know what any of these settings do, so explain each one you touch." It will translate as it goes, and you will finish the afternoon knowing more than when you started.
I made the changes, screenshotted the pages again, sent them back, and asked it to confirm they had applied the way it expected. Worth doing even when you can read the page. Router interfaces silently ignore settings more often than you would think.
The screenshot is also your undo button. Take one of every settings page before you change anything. If it goes badly you have a picture of exactly how it used to be, which is worth more than remembering.
That loop, show it the page, do what it says, show it the page again, is the thing I would hand to anyone. It works on far more than routers.

Then, and only then, I bought an extra device

The last prompt was the one that paid for the whole exercise:
Me: Based on everything you have measured and everything you saw in my settings, what is left that costs nothing? Do that first. Then tell me honestly whether there is anything worth buying, what it would fix, and where exactly it would need to go.
The normal version of this goes: wifi feels bad, read some reviews, buy a mesh system, hope. If it does not help, assume you needed the more expensive one.
I got the free changes in first, which is not what a shop would have told me.
The bigger win was knowing that the hardware purchase was backed up by Claudes diagnosis, not my gut feel that I needed another device.
What it was worth
At my desk, 15 metres from the router:
Before | After | |
|---|---|---|
Download | ~100 Mbps | ~328 Mbps |
Upload | ~70 Mbps | ~318 Mbps |
Dropouts | Several a day | None |
Next to the router it went from 422 Mbps to 1247. So roughly 3x download and 4.5x upload, in the chair I sit in every day.

Why this is a Cowork story and not a wifi story

Nothing here was secret knowledge. All of it has been on the internet for 20 years and anyone could have looked it up at any point.
I have spent 20 years being the person other people send their screenshots to. What struck me is that the thing making that useful was never really the knowledge. It was being willing to look at your specific situation instead of giving you the general answer.
So for me the wall was never vocabulary, it was effort. For most people it is both. Either way it is the same wall, and pointing at a screen and saying "tell me which one to click" now goes straight through it.
My take, and I will defend it: almost every tool we use to check whether our own technology is working belongs to the company that sold it to us, and it is pointed at their equipment rather than at our experience. For most people there was never a realistic alternative, until Claude Cowork.

Your task this week
30 minutes, and you need Cowork or something like it, meaning an assistant that can actually run things on your computer and read images, rather than just chat.
1. Sit wherever the internet annoys you most and prompt this.
My wifi is terrible. [Describe what actually annoys you.] The app says everything is fine, which does not
match what I am experiencing. Can you properly test it from this laptop, from where I am sitting right now,
rather than guessing? Explain anything you find in plain English, and do not suggest fixes until you have
measured something.2. Get up and move. Do this 3 or 4 times around your home.
I am now in [wherever]. Run exactly the same test again so we can compare.3. Open your router's settings page and screenshot every tab. Do not read them. Just send them to Claude.
Here are screenshots of my router's settings pages. I do not know what any of these settings do, so explain
each one you touch. Based on everything you measured, tell me exactly which to change and what to change
them to, and flag which changes are safe and which could knock the house offline.Make the changes, screenshot the pages again, send them back, and ask it to confirm they applied properly.
4. Then ask what is left.
What else costs nothing? Then tell me honestly whether anything is worth buying, what it would fix, and
where it would need to go.Your results will look nothing like mine, because your home and your gear are not mine. That does not matter at all. The sequence is the same: describe the problem, insist on measuring before fixing, move around and measure again, then show it the screen instead of trying to decode it.

Elon's Launching His Own "iPhone." 3 Stocks to Buy to Prepare
If you'd bought Apple before the first iPhone, you'd have seen peak gains of 7,537%. One product turned ordinary investors into millionaires.
Now Elon Musk is launching his own version — a device he claims will be "10x bigger than the largest product in history."
Target launch: July 22. Just days away.
The people who win on launches like this don't buy after the headlines. They position before. Our analyst pinpointed 3 stocks set to ride the launch — with entry guidance and price targets, plus a bonus 4th pick (the most undervalued name in the supply chain) and a 3-phase playbook showing exactly when to buy, when to add, and when to take profits.
100% free. No credit card. Over 2,500 investors have already read it.



👉️ Forward this to someone to diagnose their wifi issues. When 1 person subscribes from your link, you unlock the Beyond the Buzz Prompt Vault: 10 copy-paste prompts that do real work. Forward and unlock the Vault.

Cheers, Tim


