Skip to main content

Posts

Showing posts from 2025

Steer

Let's get this out of the way; My biggest issue with AI writing apps is they strip away the tone and character of writing. Imagine two people send you an email. They both include the phrase "Let's connect and align on this". Out of context, how would you identify who said that?  As opposed to "Let's meet for a working lunch (So we're both still paid on the clock) and go over paperwork," or "See me at 2 PM to revisit the project". Companies could save a lot of time and money on "team-building events" if the vice grip of professionalism in writing communication was beaten back. You can be professional and authentic without being inappropriate - or you can be the same as everybody else.  But meanwhile, we have many tools to help proofread your writing - Something I don't mind AI doing. Steer ( Product Hunt ) is described as ''a native macOS utility that allows users to quickly fix and iterate their writing in any app or w...

A 2-week Trial of T-Mobile Home Internet

     The Xfinity app showed usage of the past 3 months: We used less than 40% each month, for about $80 USD a month.   No thanks! That cuts into the movie budget! Before we save some money (about $15/mo), let's test how T-Mobile Internet unlimited data works for 2 weeks.    There are 15 devices for this test; Smart TVs: 4 Laptops: 4 Printer: 1 Smart Home Speakers: 3 Game Consoles: 1 Phones: 1 (There are other phones in the home but they stick with data) Other: 1 Total : 15  I made tables for 3 entries a day across 3 days to test the Xfinity service we have. Here's one;   Xfinity is pretty speedy - Download times are between 227 - 236 Mbps, Latency between 24.5 - 25.5, Jitter between 5 - 6.68, and 0 packet loss.  Let's quickly define the terms in the table;    Date/Time - The date and time of the data gathered. Download (Mpbs) - How fast your network gets data. Upload (Mbps) - How fast your network uploads data. Latency ...

A Beginner's Guide to Archiving History

 You can become a digital archivist!    If you visit a museum or national park, take photos of displays, written information, plaques that may come under contention later. If there are signs about not photographing art or exhibits, respect that  - stick to the words.  Back these up to a cloud service, or physical media that does not connect to the internet. Take clear photos.   Use archive websites to find pages removed from the current internet. The Wayback Machine is popular, and GovWayBack focuses on missing pages from government sites.   Libraries do sell old and worn copies of books; Buy some. Hold onto those with information that may come under contention. Take photos - scan them if possible. Look for online file compression tools to zip multiple files into a compressed folder. What about Google Drive?  There are concerns that "certain" files may be deleted or removed. Simply put, I wouldn't trust them to not remove a pdf upload of "Letters ...

We have to go deeper: Local PostGRES databases in Docker for development environments

  However, I keep...not being able to enter the PostGRES database via terminal when it's running in a Docker image.  So I scuppered my work, redownloaded the image, took clearer notes with commands and examples on how prompts should appear, and got inside, taking all of 10 minutes. It's like entering a multi-layered vault.  If you're starting a Docker container through the terminal, it is case-sensitive. docker exec -it RUNTCPIP-DBRep /bin/bash is different than docker exec -it RUNTCPIP-DBREP /bin/bash     Downloading Open Datasets Two Ways For Test Data Heard of Kaggle? Now you have. The first way: I took this dataset (free to use) and downloaded it to the Linux system in the container.  I updated packages with apt-get -y and installed curl -(  apt-get -y install curl). (Keep in mind you will likely have to change the path Kaggle gives. I used a period [.] and ended up putting it in a directory I could only reach via an odd song and dance of entering ...

Perplexity AI: The App, For The Everyday Person

   Perplexity AI, according to BuiltIn , is an AI-powered search engine - and it's now valued at 1 8$Billion dollars, with a B. I had it on my phone for research testing - something I do between contracts for money - and simply kept it all this time. With it making a resurgence, I can show you if it's viable for every day use cases.  I did not use it to generate "art" or writing.  Screen Reading and Photo Identification. I have used Perplexity to read Chinese characters on my screen, asking to point out the radicals, tone, and meaning of unfamiliar characters. There are minor differences between what Perplexity answers with and what Duolingo and DuChinese deal with, but I know enough Chinese* to figure out the difference - though a recent study calls the accuracy into question. For instance, below I've asked it what the radical is in  æ°´ (shui, water)      [alt: The character in your image is  æ°´ , which is the Chinese character for "wate...

LinkedIn Premium in 2025

  For a company that sells our data (If the product is free, the product is me...and you), LinkedIn offers Premium,.  One can send 5 messages for $40 USD with Premium...or use hunter.io , finding email addresses of people at companies for free. There are issues with that, but I have a free trial!  LinkedIn put forth effort to make the cost worthwhile, so let's see if the options are worth it.   How Much of LinkedIn's Revenue is Premium? I couldn't tell, but COO Daniel Shapero says in 2024, it was a 1.7$ Billion dollar business . He raves about how Premium users like the AI Writing Suggestions tool - but hold that thought.  (On the payment page, it says "Why do we need your card for a free trial? To provide a seamless subscription experience!" It's so you forget to cancel the trial. To be fair, they do say they'll warn you 3 days before it expires, but we know what it's really about)   LinkedIn Learning  One part of LinkedIn I genuinely enjoy are the ...

Windows 11

    Semi-Retro Computing Style   After I placed my start menu and icons back in their proper position , I vibe more with the visuals. They me of Web 2.0 Gloss - But you may know it by another name, Frutiger Aero - of the mid 2000s and Vista. I do miss the active tiles in the Start Menu. I liked pinning images to it Ain’t No Party Like An S-Mode Party So I took 11 out of S-Mode to install Firefox...not knowing Firefox was already in the Windows Store. I swear , I searched. Freeing myself from S-Mode was inevitable anyway - There were many third-party packages to install anyway. After all, chocolatey and Terraform are not in the store. For people who simply need to type and browse, S-Mode is your best Mode. Setting Up The Place To Code Visual Studio Code (VS Code) downloaded without issue, which is not the case for the AWS CLI through the VS Code Extensions store. The popup did helpfully have a link to where I could download it via Powershell that worked. Time to re-login...

Deploying and Managing Scalable, Cloud-Based Infrastructure in Azure

This blog post is located over here; Managing Scalable Azure Infrastructure    Here are my notes while the above portfolio is being restored; I...accidentally deleted it while mass cleaning out my Drive and didn't realize for a week. Oops. If anyone has a contact at Google who would like to help, I'm on LinkedIn.  They're messy, but the gold shines through. In Summary (Load Balancing and maintenance); I created two azure load balancers for web app redundancy and speed; this eliminates single points of failure. Health probes let me know if something needs attention.   Distributes web traffic across multiple virtual machines, making it resilient + scalable, maintain performance, and eliminate single points of failure. A health probe makes sure everything is up and operational, and LB rules distribute HTTP traffic. Hands-on experience configuring 2 azure load balancers, a tool that, at it’s base basicness, has multiple components that ensure uptime, optimize web traffi...