Internet of Bugs

35 year software professional discussing software careers, quality, bugs, and AI. Special focus on LLMs generating code, and how programming might change as AI advances.


Internet of Bugs

NOTE: Signups are closed for the moment - I'm booked for the next couple of weeks. Stay tuned.

I'm not a fan of combining "channel status" type information in the same videos as channel content, so here's a standalone video talking about the project I'm working on about helping people make their own Software as a Service projects, and how you might be able to help, if you're up for that.

6 months ago (edited) | [YT] | 25

Internet of Bugs

Just FYI - since a lot of you ended up here because of my Devin video - yes, Devin is now (at least supposedly) available. However, it's $500/month, so don't expect any review of Devin from me anytime soon.

9 months ago | [YT] | 151

Internet of Bugs

For those of you who might be interested in my ideas about the future of the channel, here's a video discussing it:

1 year ago | [YT] | 11

Internet of Bugs

I had some questions/suggestions after my last video about using specific models - usually either GPT-4 or Claude 3.5 Sonnet. As far as I know - I did.

GitHub says that Copilot is using (some iteration) of GPT4 (see github.blog/changelog/2023-11-30-github-copilot-no… )

And below is what the version of Cursor I used in this experiment was configured to use (I did not change these settings - so this would have been the default at least at some point).

I can't independently verify which models were actually being used (I didn't dump and decode the network traffic), but I have no reason to believe it's inaccurate.

1 year ago (edited) | [YT] | 100

Internet of Bugs

Here's the last segment of my interview with ‪@NobodySpecialFinance‬

I really enjoyed this interview, and if you did, too, be sure and drop a comment on these videos so we'll know. If enough people want us to, hopefully we'll figure out some way to collaborate again.

1 year ago | [YT] | 13

Internet of Bugs

Here's part two of my interview with ‪@NobodySpecialFinance‬

Hope you enjoy it.

1 year ago | [YT] | 21

Internet of Bugs

Hey, folks!

Here's part one of an Interview I did with ‪@NobodySpecialFinance‬ about AI Demos.

If you haven't checked out his channel - and you're interested in the financial incentives behind this hype (which you absolutely ought to be) - you should check out his channel. Especially his series on "AI Mania." He's done some great work there.

1 year ago | [YT] | 49

Internet of Bugs

Ok, I have a question for the community. This has come up in a few comments, and-although I'm pretty sure they were just trying to be funny and sarcastic-the underlying issue is real.

Just as a thought experiment - what would it take (short of actually meeting up in meatspace with you or someone you know) for me to prove to you I'm a real person?

Alternately, in the abstract, in this day and age, if someone hypothetically started a YouTube channel in 2024, and purported to be talking live to camera, and had a bunch of accounts (GitHub, LinkedIn, etc) that they pointed to and claimed to be theirs, but you couldn't meet them in person for logistical reasons, what would constitute proof they weren't an AI generated Avatar?

I'm not asking what would *convince* you - I'm asking: what could you be sure of? (Ignoring "what if this is a computer simulation?" and "how do we really know anything is real?" kinds of philosophical questions.)

I don't really care if people think I'm an AI or not, but proving who is actually real and who isn't is something we're going to need to able to do. So how?

I'm trying to think of what couldn't be faked, and honestly, I'm having a hard time coming up with something.

P.S. And, just for context - watch this: www.microsoft.com/en-us/research/project/vasa-1/
I didn't see that until after I'd posted this video, although I kind of implicitly assumed it would be possible.

1 year ago (edited) | [YT] | 86

Internet of Bugs

# Video Topic Backlog

Hi, Subscribers (and hopefully future subscribers)!

First off, thanks for helping this channel get off the ground. I'm blown away by how much traction it has gotten in 6 weeks with only 12 videos.

I thought it might be fun to give you all a peek behind the curtain. This is a snapshot of my current video topic backlog - the videos ideas that occurred to me (or were suggested by one of you in the comments - thanks for that!) that I wasn't going to start work on immediately, so I put them here so I don't forget them. Some of these ideas have been done, now, - those are the ones with the solid circle to the side of them, like "Adversarial AI Attacks" on the top-left, which was on of the 5 bullet points in my "5 new Software Engineer Jobs that 'AI' will make" video from March 22nd.

Not all the videos I'm currently working on are in here. For example, the two videos that were cut out of the "Interviewing is Broken" video–that are tentatively titled "The Myth of the 10x Developer" and "If not LEET, then what?"–went straight from the Interviewing is broken outline to their own outline files without going through the backlog.

These are in no particular order - which is why they're in Mind-Map form. I find when I put things in an outline, my brain automatically wants to default to the ones at the top, so if I don't want to force an implied sequencing, I use MindMapping to record things.

I'm not sure how much you'll be able to make sense of my terse topics, but hopefully you'll be able to match them up with that video drops.

No promises exactly when (or if) I'll get to any specific one of these - I don't want to do a video until I've thought through what I want to say and have convinced myself it will be worth other people's time - and that can be unpredictable.

Although I seem to have enough material for roughly a year or so if I average one a week, I'm always curious about what people want to see, so if you have ideas for videos you'd like to watch, drop me a comment, and if any of these seem especially important, exciting or relevant to you, let me know, and if something turns out to be more popular than I suspect, I'll try to get it done sooner.

Thanks, sincerely, to all of you,

-Carl

P.S. I also have another mind map that's about 3 times as long for a book I was shopping around on Software Testing an Quality - the foreword of which became the majority of the "How Testing Software will need to change in an AI-driven World" video from April 2nd. So expect videos about that kind of thing, too (although that video wasn't nearly as popular as I'd hoped/expected, so I'm not sure how much of that will be worth doing).

1 year ago (edited) | [YT] | 178