Jon
banner
musselwhite.dev
Jon
@musselwhite.dev
Computer science PhD student at Howard University, homelab tinkerer, digital photographer, and Oxford comma adherent.

📷 joncaptureslight.com
👨‍💻 musselwhite.dev
Yes. Any backup strategy is better than no backup strategy. Improvements can always be made, like adding automation, remote storage, multiple storage mediums, time-based snapshots, etc., but perfection is the enemy of progress. Something is better than nothing.
is this a good backup strategy
November 13, 2025 at 2:47 PM
Shout out to Mountaineer Gas for requiring new customers to be present at the location all day because they won't require their technicians to call in advance no matter how frigidly cold it is there at the vacant house.
November 11, 2025 at 6:15 PM
Reviewing a product on Costco's website feels applying for a credit card at a credit union. I don't know how to explain it. Why does it need my member number when I'm logged in??
October 31, 2025 at 12:22 PM
Google Pay doesn't work on GrapheneOS because it apparently offers too much security by adding the ability to block network and sensors. Very odd.
October 31, 2025 at 10:59 AM
Reposted by Jon
"We’ve reached out to Google to inquire about why a custom ROM created by volunteers [@grapheneos.org] is more resistant to industrial phone hacking than the official Pixel OS. We’ll update this article if Google has anything to say." arstechnica.com/gadgets/2025...
Leaker reveals which Pixels are vulnerable to Cellebrite phone hacking
Cellebrite can apparently extract data from most Pixel phones, unless they’re running GrapheneOS.
arstechnica.com
October 31, 2025 at 12:14 AM
This video makes a good point about LLM usage in automations. When the desired result has a single well-defined correct answer, write some actual code to do it (or have an LLM write it). It's cheaper and takes the RNG out of the loop.
w3cj.com CJ @w3cj.com · 23d
Some devs are leaning a bit too hard on AI for solutions and workflow.

In my latest video I show a few examples and give some advice on what to do / use instead.

Watch 🍿 www.youtube.com/watch?v=pVBK...
Stop Using AI For This
YouTube video by Syntax
www.youtube.com
October 30, 2025 at 1:20 PM
This is why I advise organizations to containerize all apps for Kubernetes instead of making them with a cloud service's proprietary tools. That way, you won't be dependent on a single cloud provider when they experience an outage or if you need to pivot for any other reason.
timburga.com Tim @timburga.com · Oct 20
1980: the internet routing system provides a virtually unlimited number of paths between any two machines to ensure maximum connectivity and resiliency

2025: can't flush the toilet because AWS East is down
October 20, 2025 at 5:30 PM
Recently, I reposted a link to a blog post describing an MIT study that showed an accumulation of cognitive debt when using AI writing assistants. This morning, I discovered the blog is also full of pseudoscience, so I have deleted those links and am sharing a direct link to the study instead.
Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task
This study explores the neural and behavioral consequences of LLM-assisted essay writing. Participants were divided into three groups: LLM, Search Engine, and Brain-only (no tools). Each completed thr...
arxiv.org
October 20, 2025 at 3:08 PM
This is an exciting development! A huge challenge in software freedom on mobile devices is the lack of open source drivers. Imagine being able to install a Linux-based distribution of choice on a mobile device and having it just work instead of being limited to whichever distro supports your phone.
October 15, 2025 at 8:43 PM
Reposted by Jon
it's kinda weird that all the software i am expected to use for work are all written by distributed teams, go, python, postgres, linux, chrome, k8s etc

and despite being told "the best teams work in an office together" i don't know of any software i use that's actually written that way
October 13, 2025 at 5:00 PM
It's too late. They showed their hand. It's not wise to buy gear from a company that is trying to find ways to limit consumer choice. They'll just try again later.
October 9, 2025 at 4:42 AM
The concept of a total loss on a comprehensive car insurance policy is terribly wasteful because it silently diminishes the value of the policy every year even as rates stay the same while the replacement cost rises, leading to situations where the car is "totaled" over cosmetic damage.
September 30, 2025 at 10:25 PM
This is a good reminder that LLM interfaces use a random number generator to select the next token in a stream from a short list of most likely options based on data observed during training. That's not intelligence. That's dumb luck, no matter how lucky it can seem.
Lmao 🤣
September 22, 2025 at 4:48 PM
Loan scammers adding insult to injury in their robocall by saying "this may be the last chance to wrap this up" as if they'll actually stop calling about a "loan" that is in the "final stages of being approved." ...
September 15, 2025 at 4:41 PM
Tuesdays are the most annoying day to be on T-Mobile. They dropped their Scam Shield app and put that feature in their "T-Life" app, which runs much slower on Tuesdays -- the days when ad sales are more important than customer experience. No, I don't want a FREE drink (with $8 purchase), thanks T.
September 9, 2025 at 6:58 PM
I wouldn't tell anyone if I won a big lottery jackpot, but there would be signs.
DGX SuperPOD
Built for the Unique Demands of AI.
www.nvidia.com
September 9, 2025 at 5:14 PM
Here's neat trick: GitHub public keys can be downloaded from `https://github.com/<username>.keys`. The file lacks comments, but it is properly structured to be dropped into `~/.ssh/authorized_users` when deploying an environment with an SSH server, reducing the risk of deploying stale credentials.
September 9, 2025 at 2:51 AM
Are any studies being done on the impact of early warnings being too early for the general population to accept? For example, if a tsunami could be predicted a year early, would that give energy to a counter movement of people who reject science so they can get louder than the truth?
September 7, 2025 at 3:15 PM
I've been on the verge of switching back to Linux for my base OS for a while, but some of my apps rely on the GPU and don't have good Linux alternatives. If I pass the discrete card into a Windows VM, I wouldn't be able to use it to drive my three monitors on the host. Does anyone have a solution?
September 6, 2025 at 5:12 PM
Reposted by Jon
I feel like one of the many effects of LLMs on our industry is going to be making it really clear that the coding was never the hard part.
September 5, 2025 at 8:53 PM
If Aurora can't figure out how to deploy a reflective triangle warning device 100 feet up the road without a human driver when their big robot breaks down on a public road, maybe they shouldn't send robots on the road without a human...
September 5, 2025 at 1:54 PM
The co-principal investigator on that was one of my professors! Dr. Gloria Washington broadened my understanding of human computer interaction and affective computing. It's so cool to see this project pop up in the news.
Cool story in @washingtonian.com about @howard.edu's Lucretia Williams: "top-tier voice-interpretation systems have higher error rates for Black users than for white ones—a 22 percent gap for Apple, 15 percent for Amazon, and 12 percent for Google."

www.washingtonian.com/2025/08/05/h...
How Howard University Is Helping Tech Understand Black Speech
Howard and Google teamed up to improve voice recognition.
www.washingtonian.com
September 4, 2025 at 4:16 PM
Reposted by Jon
Results so stunningly clear they inspired this classic xkcd (xkcd.com/2400/):
September 1, 2025 at 7:32 PM
Reposted by Jon
I vaguely remember tweeting something like the country that effectively captures this free energy source at scale will rule the 21st century. whelp.
August 29, 2025 at 8:05 PM
TIL how toothless the CAN-SPAM act really is. Shoe Carnival won't stop sending me notifications about a $10 reward as an excuse to advertise their inventory, but only the federal government can sue them over the CAN-SPAM act violations (or maybe an ISP if it can prove damages).
Do I have ANY recourse under the CAN-SPAM Act?
www.law.cornell.edu
August 27, 2025 at 2:39 PM