Bob Flynn

@bobjinx

Cartoonist. Character Designer. Doodler. Director of Art & Animation at FableVision Studios jinxthemonkey.com

326 Following3,798 Followers1,179 Posts

YouTube YouTube Bluesky

amateur

Is this your profile?Claim it and bobjinx.dirty.social is yours.Claim profile

Bob Flynn@bobjinx· 19d

It’s hard to tell from the article if this was a genuine accident (both companies benefit from the PR if any of the story has been exaggerated). One day we’re gonna wake up to the news that a significant amount of the internet has been wiped out, or corrupted. www.nytimes.com/2026/07/21/t...

OpenAl Says Its A.I. Models Went Rogue and Attacked a Digital Library
The incident, which targeted the computer systems of another company called Hugging Face, happened while OpenAl was testing the systems.

Bob Flynn@bobjinx· Jul 6

I created the poster art for the show, titled 'A Matter of Scale' (the final list of exhibitors is may change, but this is the latest)

Bob Flynn@bobjinx· Jun 29

For real. I've been building a castle with my son (our own design), so I know a little too much about this sort of construction. But the part use is indeed pretty clever. It's a mix of jumper plates on the base and gray wall panels. Instructions if you're curious! www.lego.com/cdn/product-...

Bob Flynn@bobjinx· Jun 25

Every flower counts.

Bob Flynn@bobjinx· Jun 11

Currently awaiting:

Bob Flynn@bobjinx· Jun 8

Now thanks to @ani-obsessive.bsky.social , there's a gif of this very moment! Featured in their most recent newsletter "Slow Stories Don't Have to Be Boring."

Bob Flynn@bobjinx· Jun 4

Wow, if these statements really sum up the moral conundrum, don’t they. By giving up on decision-making, you’re sidestepping responsibility. Additionally sidestepping whatever mental rigor it takes to make a decision. Moral, creative, or otherwise.

If a company builds a machine that, when fed descriptions of assorted ethical dilemmas, emits sentences either of the form
"Compromise your values" or "Don't compromise your values," it is not building a tool that assists people in their decision making; it is encouraging people to stop making decisions. The writer L. M. Sacasas has said, "Our technological systems, by nature of their design and the ideology that sustains them, are machines for the evasion of moral responsibility." He was talking about social-media plattorms, but his observation is, if anything, even more applicable to LLMs. Whenever a person delegates a decision to an LLM, they are trying to off-load accountability for that decision, and if a company that sells an

making decisions. The writer L. M. Sacasas has said, "Our technological systems, by nature of their design and the ideology that sustains them, are machines for the evasion of moral responsibility." He was talking about social-media platforms, but his observation is, if any hing, even more applicable to LLMs. Whenever a person delegates a decision to an LLM, they are trying to off-load accountability for that decision, and if a company that sells an LLM portrays the product as having a moral center, it is offering a way for its customers to abdicate their responsibilities.

Bob Flynn@bobjinx· May 31

elf-like hero in green cloak / #LegoMOC

Bob Flynn@bobjinx· May 28

I realize this has always been the case. Execs and PR not understanding how things work. But how does something so blatantly false slip through the cracks? So easy to check. So easy to understand how Xerox was used. (response from Buzzfeed via Cartoon Brew: www.cartoonbrew.com/artificial-i...)

We would have loved for Loryn to be involved as one of the people who worked to bring Cuppy to life while employed full time at BuzzFeed. We shared with her that human creativity would remain at the core of this project, with writing, storytelling, and animation being developed by humans and AI being used as a creation tool to help facilitate that.
We explained that this was similar to Walt Disney embracing Xerox technology to make "inbetweening" easier - and that our partners at Amazon MGM Studios share our ethos, with a commitment to ensuring human creativity remains at the center of everything they make.
However, she made it clear that she was categorically opposed to the use of Al in all its forms, and we respected that she did not want to be involved as a result.
That is absolutely her right. But her personal opposition to AI cannot determine how BuzzFeed develops IP that it owns, or deny the many other talented creators involved in this project the opportunity to do their work.

Bob Flynn@bobjinx· May 27

from the paper: "First, these models are autoregressive systems optimized to select the most probable next token, which fundamentally favors predictable and conventional outputs over novel or unexpected ones." "Second, they undergo instruction tuning...to produce "safe" or aligned responses."

One might expect the tendency of LLMs toward homogenization to decrease as GPT models and other LLMs continue to evolve. However, previous research suggests that LLMs share two systemic factors that inherently limit their output diversity.
First, these models are autoregressive systems optimized to select the most probable next token, which fundamentally favors predictable and conventional outputs over novel or unexpected ones (Brown et al., 2020). Second, they undergo instruction tuning with Reinforcement Learning from Human Feedback (RLHF) to produce "safe" or aligned responses (Kirk et al., 2024; Mohammadi, 2024). These processes, while improving fidelity and reducing hallucinations, can inadvertently reduce creativity and encourage homogenized responses. Indeed, our findings showed that a newer version of GPT-4 (gpt-4-0125-preview) actually yielded higher homogenization than the previous model (gpt-4-0613). Further supporting this tendency, adjusting parameters such as 'presence_penalty' and 'frequency penalty' increased semantic diversity and decreased homogenization but at the expense of coherence and readability (see Table S10 for details). Even with these modifications, Al-generated essays still had a lower diversity growth rate than those written by humans, suggesting that homogenization may be an inherent feature of current LLMs.

Bob Flynn@bobjinx· May 27

Been hearing people justifying the sludge machine at the concept and brainstorming phase of a project. "Human-written essays offered up to eight times more new ideas than those produced by A.I." AI doesn't come up with new things. It's a plagiarism/resampler. www.nytimes.com/2026/05/27/o...

In a separate study, the team found that human-written essays offered up to eight times more new ideas than those produced by
A.I.
Another experiment run by a different research team compared short stories written by humans to those written with A.I. assistance. As with the student essays in Dr. Green's study, A.I.-assisted essays had more interesting vocabulary and were rated more enjoyable to read, but the underlying story lines were more homogenous. Distinctive and offbeat ideas — with surprising characters or unusual settings — are often shunted to the side when A.I. is involved.

Bob Flynn@bobjinx· May 7

I thought this line was interesting (not that I’m suggesting any AI could replace a good teacher). Those who used it to get clarification is similar to a kid raising their hand in class. A teacher would never provide the answer. That’s where the student-chatbot dynamic is so damaging.

Effects Are Concentrated Among Users Who Seek
Direct Solutions
Analyzing self-reported Al usage patterns, we find that the majority of participants (61%) used Al to get answers directly. These participants showed the largest declines in performance and persistence - not only compared to control participants but also compared to participants who used AI for hints or clarifications. Participants who used
AI for hints showed no significant
impairments relative to control.

Performance and persistence declines are concentrated among participants who obtained direct solutions from AI. (a) Al usage groups show no significant differences in solve rate or skip rate at pretest (one-way ANOVA), suggesting comparable

Bob Flynn@bobjinx· Apr 26

Surprising how little Merlin's beard factors into any of the animals below (it doesn't!). Blue does most of the work—the glasses, brows, and sometimes the mustache. Color model sketches by Milt Kahl (I made a composite of them, boosted the levels) Source: andreasdeja.blogspot.com/2012/05/wiza...

Bob Flynn@bobjinx· Apr 26

A good character design exercise would be transforming your character into a bunch of animals, a la Merlin.

Bob Flynn@bobjinx· Apr 26

Storyboard sketch (Bill Peet?) compared to a final frame. A lot of the drawing's energy and personality was captured in the final, I'd say. Disney's 'The Sword in the Stone' (1963) Source: characterdesignreferences.com/art-of-anima...

Bob Flynn@bobjinx· Apr 25

Is there such a thing as too many creatures in Star Wars? Is it better for world building to scrape every puppet and creature suit from the Mos Eisley Cantina? Do we dig a bunch of Salacious B. Crumbs squawking from the trees in The Mandalorian? Or do we prefer it as a singular Gonzo of a muppet?

Bob Flynn@bobjinx· Apr 25

This almost became a rambling post on Ithorians, the Talz species, and whatever these guys are, but they all get to the same conundrum. They were imprinted on me in the Cantina. (it distracts me) Writers have a desire to have Star Wars feel real and make sense. But why not design new creatures?

Bob Flynn@bobjinx· Apr 25

Another fan favorite has been since retconned into a species called Rodians. In contrast to Greedo the criminal, these are sympathetic characters, including a baby and a Jedi! But my old-timer brain already has the design coded as the face of the alien Han Solo shot underneath the table.

Bob Flynn@bobjinx· Apr 25

Star Wars has been cloning itself at least since the prequels. The fan favorite bounty hunter Boba Fett (and his ship) got a face, a dad (Jango), and a duplicate ship. Later stories spun off that cool armor into an entire culture & planet (Mandalorians/Mandalore) complete with a religious sect.

Bob Flynn@bobjinx· Apr 25

I’ve been watching the Clone Wars animated series for the first time with my son. I like it. But one character element that gets my brain going is how they opted to make entire species out of what I’d always considered great one-off creature puppet designs. Created for establishing shots.

Bob Flynn@bobjinx· Apr 23

A mustached Lemongrab spotted at the American Museum of Natural History in NY.

Bob Flynn@bobjinx· Apr 13

As a recovering perfectionist, this rings true.

She must be squeezing him hard.
How can I draw that?

Must be somewhere...
The lines I'm looking for.

There's a correct placement for those lines.

If I can find them, I can show the "squeeze."

Bob Flynn@bobjinx· Apr 9

Search for 'diagram of a guitar' and you get something like the left image. Pick a book from my kid's library stack and you get the right. (I now know to grab said book or search 'The Way Things Work' for most mechanical questions I have)

Bob Flynn@bobjinx· Mar 30

Legitimate advances in speed, compromised by crazy expensive infrastructure build out, bolstered by irrational investment. www.smithsonianmag.com/history/robb...

Transcontinental routes enabled people and goods to traverse vast distances at what were then considered lightning speeds and drastically reduced costs. As economist Arthur Twining Hadley wrote in 1885, "Two generations ago, the expense of cartage was such that wheat had to be consumed within 200 miles of where it was grown. Today, the wheat of Dakota, the wheat of Russia and the wheat of India come into direct competition. The supply at Odessa is an element in determining the price in Chicago."

A railroad of any size required an enormous outlay of capital to purchase locomotives, rolling stock, stations, terminals, train yards, and other equipment and infrastructure. As a result, the rise of the railroads inspired entirely new forms of corporate securities and governance. They also yielded stupendous levels of fraud and chicanery, wrote business historian Alfred D. Chandler Jr. in his 1965 book, The Railroads: The Nation's First Big Business. Railroad owners often made more money on construction, land and mineral rights than they did on operating the tracks themselves.

Cooke later harnessed his pioneering direct-to-consumer investment concept to offload bonds in a scheme selling the latest technology-railroads —to the next generation of individual investors. "It was like selling to crypto investors today," says historian Richard White, author of Railroaded: The Transcontinentals and the Making of Modern America. "These investors were basically living off of the promises of famous people they had trusted before."

Bob Flynn@bobjinx· Mar 21

I’ll be at the Ashland Library in Massachusetts this afternoon doing a short workshop on Doodling Dinosaurs. It’s a fun comics event for all ages, and includes kids tabling with their own comics. www.bostoncomicarts.org/ALYAC

1:45 PM - 2:30 PM
Doodling Dinosaurs with Bob Flynn in Memory of Our Friend Dan Moynihan)
Children's Storytime Room - Let go of your expectations of good or bad drawings by learning to play while you draw. In this workshop we'll ditch pencils and erasers and free-draw with markers. Start with a shape and a squiggle and keep building onto it. We'll be making up creatures and dinosaurs.

Bob Flynn@bobjinx· Mar 5

Just don’t, whatever you do, use it to file your taxes.

To assess the technology’s ability to file a federal income tax return, The New York Times tested four A.I. chatbots — Google’s Gemini, OpenAI’s ChatGPT, Anthropic’s Claude and xAI’s Grok — to see how well they fared with eight fictional tax situations written as part of training materials by TaxSlayer, a tax-filing service.

They struggled, hard, miscalculating the refund or amount owed to the Internal Revenue Service by an average of more than $2,000. Even when provided with all the necessary materials, including all the forms they needed to fill out, the chatbots whiffed on some calculations.

Bob Flynn@bobjinx· Mar 5

Taxes aren't all that hard to do. Why would you have it do anything of *any* importance for you?

NYTimes headline: A Word to the Wise: Don't Trust A.l. to
File Your Taxes

Bob Flynn@bobjinx· Feb 28

Night at the Pocket Monster museum /

Bob Flynn@bobjinx· Feb 25

A lot of the best character sheets (and a lot of the best character directors) spend just as much time on the X's of what to avoid. Often comes from giving repeated notes to artists when they draw things a little off. Best example in the set is this box here. All the ways *not to draw the mouth.