Matthew Thorne

DevOps Engineer @amazon

[Matthew Thorne @mattthorne]

I recently worked on a strange issue in a Kubernetes cluster, idle pods being throttled almost constantly. Prometheus metrics are not as simple as I thought.

https://t.co/tH7TpQ4Iu9

[Matthew Thorne @mattthorne]

I recently migrated an S3 bucket and ended up 17 million objects short.

No errors or obvious issue, just a strange gap.

Turns out, S3 hangs on to a lot more than I thought.

https://t.co/E8A3nMnwmv

[Matthew Thorne @mattthorne]

A few weeks ago, I helped a client cut over $5,000 a year from their AWS bill in just a 60-minute call.

[Matthew Thorne @mattthorne]

@davefaliskie I normally do IOS first. After I add any major dependency I test on Android though to check it still builds! I once spent 3 days trying to resolve why an app would build on IOS and not android.

[Matthew Thorne @mattthorne]

Helped a client speed up their database queries by 100x by fixing a critical performance issue.

Their queries were slowing down as data grew. After redesigning their keys & indexing , queries that took nearly a minute now run in under a second.

[Matthew Thorne @mattthorne]

@NorDotShell Vim is great... once you figure out how to exit

[Matthew Thorne @mattthorne]

@martyamark https://t.co/SQL7dLiMEj

[Matthew Thorne @mattthorne]

@sgshowman Maybe using the genetic mode is the key. I’ll give it a go.

[Matthew Thorne @mattthorne]

Has anyone found a good LLM for mobile app development?

I’ve found for anything complex in Flutter, they’ve been terrible. Great for web apps and scripts, but for mobile, they often feel like a hindrance.

Is it due to limited training data—most mobile apps being closed source?

[Matthew Thorne @mattthorne]

@ChinuKabi I’ll give that a go.

Yeah if you know exactly what you want it’s great, effectively a really good auto complete.

[Matthew Thorne @mattthorne]

@FcoPardoBaeza Definitely what I have found!

[Matthew Thorne @mattthorne]

@winterthedev I just ask the AI to search stackoverflow for any answers if it doesn’t get it right away. Rarely actually go on the site anymore.

[Matthew Thorne @mattthorne]

@khant_dev At the beginning of a project it’s pretty low. But as the project goes on it probably peaks at like 80%, generating all the boiler plate. Then drops off again as I get more into the weeds.

[Matthew Thorne @mattthorne]

Local community detection offers a scalable alternative over global network analysis, yet few libraries provide ready-to-use algorithms.

To address this, I’ve implemented Clauset’s LCD algorithm in NetworkX, enabling LCD with a single line of code.

https://t.co/OIN5edx8MF

[Matthew Thorne @mattthorne]

In December I had fun working on a project for a live event in Portugal. We developed a website that let attendees submit survey responses, with AI identifying common topics in real-time and visualising them as a dynamic network graph, showing the connections between topics.

[David 🇬🇧🇺🇦 @LondonCabbie]

London Taxis last more miles than any other car https://t.co/cOeSuV0ggI

[Matthew Thorne @mattthorne]

A car is nearly 5% more likely to pass its MOT test on Christmas Eve!

@DriveWrite @ukmotortalk @car_witter @matty_prior

[AutoPredict @AutoPredictHQ]

Your car is most likely to pass its MOT test on Christmas Eve!

#MachineLearning #Christmas

https://t.co/MlEZiihVr7

[Matthew Thorne @mattthorne]

Find out how the colour of your car effects how long it will last. If you have a green car you are in luck! @matty_prior @car_witter @DriveWrite @_HonestJohn

#AI #Automotive #green

[AutoPredict @AutoPredictHQ]

It turns out green cars last the longest! Checkout the data in our latest post below.

https://t.co/GjevjONxCN

[AutoPredict @AutoPredictHQ]

This week we look at the effect of engine size on how long a car lasts.
#MachineLearning #Automotive

https://t.co/SRBRDnlQM4

[Matthew Thorne @mattthorne]

Checkout this weeks AutoPredict post. LPG cars on average last the most miles. @DriveWrite @ukmotortalk @car_witter @matty_prior
#Automotive #MachineLearning

[AutoPredict @AutoPredictHQ]

In this weeks post we look at how fuel type effects how many miles a car does. Surprisingly LPG cars came out on top!

#MachineLearning #Automotive #lpg

https://t.co/4nHZuDQONg