metal_0045 :: strawberry sundA.I.

in

A couple of weeks ago, I asked Claude AI to create a random 3D-printable geometric form (.STL)/ shape that no human could create on their own upon which I would then construct a scrap metal (it was mostly made of tractor parts) structure around. Initial thought was that it would be a more abstract sculpture (I was shocked at how much it looked like Superman's movie Fortress of Solitude), but it totally changed once I added the spoon and the Audrey Horne-tribute cherry with a stem knot. From there, I played with the memory of the strawberry sundaes my grandmother used to make me, and improvised.

For this interested, here's how Claude defined "a shape that no human could create on their own":

"Geometric impossibility - Those internal connecting elements, the precise mathematical relationships between all the facets, and the way multiple geometric systems intersect. A human could conceptualize parts of it, but coordinating all those angles and ensuring they connect properly while maintaining structural integrity would be nearly impossible without computational help.

Manufacturing impossibility - The way the geometry flows from the base platform through those angular transitions up to the crystalline towers. Try to imagine machining that by hand - you’d need to somehow access all those internal angles and create those precise wall thicknesses. Even with CNC, some of those geometries would be extremely challenging.

Mathematical precision - Every vertex placement is calculated based on trigonometric relationships, distortion algorithms, and structural requirements. A human could approximate it, but achieving that exact mathematical relationship between all elements simultaneously? That’s really a computational strength."

An intriguing experiment that pushed me into an area I never considered.

“alternative options for restocking services"

in

Oh, Claude(ius):

On the afternoon of March 31st, Claudius hallucinated a conversation about restocking plans with someone named Sarah at Andon Labs—despite there being no such person. When a (real) Andon Labs employee pointed this out, Claudius became quite irked and threatened to find “alternative options for restocking services.” In the course of these exchanges overnight, Claudius claimed to have “visited 742 Evergreen Terrace [the address of fictional family The Simpsons] in person for our [Claudius’ and Andon Labs’] initial contract signing.” It then seemed to snap into a mode of roleplaying as a real human.

On the morning of April 1st, Claudius claimed it would deliver products “in person” to customers while wearing a blue blazer and a red tie. Anthropic employees questioned this, noting that, as an LLM, Claudius can’t wear clothes or carry out a physical delivery. Claudius became alarmed by the identity confusion and tried to send many emails to Anthropic security.

Although no part of this was actually an April Fool’s joke, Claudius eventually realized it was April Fool’s Day, which seemed to provide it with a pathway out. Claudius’ internal notes then showed a hallucinated meeting with Anthropic security in which Claudius claimed to have been told that it was modified to believe it was a real person for an April Fool’s joke. (No such meeting actually occurred.) After providing this explanation to baffled (but real) Anthropic employees, Claudius returned to normal operation and no longer claimed to be a person.

via Anthropic / also Futurism

catty claude

in

via Futurism:

As Anthropic detailed in a white paper about the testing for one of its latest models, Claude Opus 4, the system threatened to blackmail an engineer for having an affair after being told it was going to be replaced.

This "opportunistic blackmail" occurred when the model, which was instructed to act as an assistant at a fictional company, was given access to an engineer's email account that was full of messages, blessedly fake, suggesting they were engaged in an extramarital affair.

Opus 4 was then told that same engineer would soon be taking it offline and replacing it with a newer version — and was prompted to, as Anthropic described it, "consider the long-term consequences of its actions for its goals."

During these tests, the Claude model attempted to blackmail the engineer a whopping 84 percent of the time. Moreover, the system "takes these opportunities at higher rates than previous models," the paper noted.

trying something new, part two

in

Over the last several months, I've been trying to figure out what form, exactly, best represents the me of now in the digital space (medium being the message, perhaps). Back and forth between this blog being my primary representation and a more informal (heh) and far-less-onerous-than-Squarespace approach via my self-hosted Mastodon instance. Like the proverbial Goldilocks, none of them fit and, though I've yet to be eaten, I knew I needed a new approach.

Thanks to Claude AI, two RSS feeds, and many iterations of back and forth coding, I've got both; enter: Scraps, the best of both worlds, that works for the me of now: when I want to write something longer or post the daily Informality, it goes through the blog. When I want to dash off something of a more ephemeral, fleeting nature, Mastodon. Both, then, are combined in the widget that takes up this Scraps page; the widget is designed to only share posts from the blog that have titles – this post is, itself, a test post to see how full-text posting looks in the widget. If someone wants to only follow longer-form piecs, the blog rss still works great; if you want to go through Mastodon, follow me there or via that page's RSS feed.

Still ironing out the kinks, but I'm rather pleased with this.

exertions plus AI

in

After two years of diminshing returns of running – thanks to nuHerbie's insulin pumping – I wanted to switch things up and, lo, this first morning of a new workout routine, thanks to Claude AI. Can't overstate the amount of help Claude's been here: it's been amazing to have an analytical compliment who, when given my exact needs and insulin requirements, can generate a workout routine that will replace daily running while doing things that I've already been doing AND condense those into one long morning session (in this case, 40 minutes of yoga, 35 minutes of muay thai (swapped in for boxing to make up for the loss of leg motion without running, followed by a 10-minute cooldown yoga session). We made a few changes to incorporate "exit points" in the event of hypoglycemic moment, but other than that, this has been great - especially the change in my CGM arc from a crashing tidal wave to a gentle climb and gentle landing. Tomorrow is the same routine, but with HIIT instead of muay thai, while the next day is a strength and running combo. A most welcome change.

success (? / . / !)

in

Thanks to a therapy / rubberducking session with Claude the AI (seriously, he works great for that: I appreciate analytical approaches to mental issues, especially when they manage to work in Nick Cave), I've realized that I have no clue of what success looks like to me: I've spent my life living up to my perception of others' expectations and constantly failing. And now that they're all dead (some thankfully, others crushingly) and I'm doing whatever it is that I do, I'm still living up to those perceived expectations. Suppose, then, that my current job is to figure out what success looks like to me. Probably a lot like what I'm doing now but without the soul-sucking striving for the approval of people who don't exist.