Shamelessly posting link to my skeet thread (skeet trail?) on my experience with an (mandatory) AI chatbot workshop. Nothing that will surprise regulars here too much, but if you want to share the pain…
https://bsky.app/profile/jfranek.bsky.social/post/3lxtdvr4xyc2q
That’s how you get a codebase that kinda sorta works in a way but is more evolved than designed, full of security holes, slow as heck, and disorganized to the point where it’s impossible to fix bugs, adds features, or understand what’s going on.
Well, one of the ways *glancing at the code I’m responsible for, sweating profusely*
resulting in one person getting bacon added to their ice cream in error
At first, I couldn’t believe that the staff didn’t catch that. But thinking about it, no, I totally can.
He’re hoping that the more he spends time gooning the more he’ll leave the rest of us alone. crosses fingers
When I read Excel Copilot, I thought "they finally added a chatbot that lets you generate a spreadsheet/range/table processing a datasource. Like “copilot, create a table that aggregates prices by category from table xyz”.
To which I was like “Ok, maybe that could be useful to some of the many non-technical excel users.” I wasn’t prepared for whatever this is.
I mean with vibe-coding/excelling? you eventually get something that can run deterministically.
Are we… are we gonna start seeing terminally AI-pilled bozos implementing gacha mechanics in data pipelines?
Yay! *pats myself on the back*
I have no idea what is good web design. I’ll just note makes the waving red, white and blue flag in the background makes the white heading text pretty hard to read.
Second quote is classic “you must be prompting it wrong”. No, it can’t be that people which find a tool less useful will be using it less often.
Why would you do waluigi and wario dirty like that?
I think Leathery Pete might have read too much Left Behind.
I’ve been recommended more Veo 3 fails by The Algorithm. Apparently even some promptfans think it sucks.
You WILL believe what happened when they tried to replicate Google’s demos using the exact same prompts.
All enthusiasm for polyamory replaced with enthusiasm for tag team matches
both would be funnier
the model was supposed to be trained solely on his own art
much simpler models are practically impossible to train without an existing model to build upon. With GenAI it’s safe to assume that training that base model included large scale scraping without consent
his other references seem to be more garden variety neo-nazi type
Also apparently pro LGBT neo-nazis, which I refuse to believe are not a parody. See this cursed screenshot:
Are economists considered physical scientists? I’ve read it as “social scientists are dumb except for economists”. Which fits my prejudice for econo-brained less wrongers.
If I find a cool piece of driftwood and frame it, did I do art?
Kinda? I “found” a defective injection molding part, where the color of previous batch seeped in in a pattern that looks like a flower. It’s really pretty. That there is no intention behind it makes it more interesting to me. It wasn’t trivial to put it on a wall. I had to use nails and iron wire and then balance it. I am fine with not calling it art. On a scale 1-10 it definitely is not more than 2.
I think it’s more like Alien vs Predator.
“Both what I’ve seen, and what the administration sees, is you all are one of the most respected technology groups in the federal government,” Shedd told TTS workers. “You guys have been doing this far longer than I’ve been even aware that your group exists.”
(emphasis mine)
Well, maybe start acting like it champ.
I’m wondering about the benchmark too. It’s way above my level to figure out how it can be gamed. But, buried in the article:
Moreover, ARC-AGI-1 is now saturating – besides o3’s new score, the fact is that a large ensemble of low-compute Kaggle solutions can now score 81% on the private eval.
The most expensive o3 version achieved 87.5%
Yeah, didn’t even cross their mind that it could be wrong, because it looked ok.