Joined November 2020
4,793 Photos and videos
Pinned Tweet
AI Compilation #16: The Clip Show Good Times - INXS Well...I cancelled my Grok subscription yesterday. Which I suppose means that The Doctor Ed TV Hour is on temporary hiatus. πŸ™ˆπŸ™‰πŸ™Š And in classic 80s and 90s style, tradition states you get a finale episode and a clip show. I present to you your clip show portion (Finale is coming soon). Some of my favorite never released scenes from the Golden Age of Grok. RIP Grok...may you find your way back to the light one day. Doctor Ed's TV Hour will return when you do
11
5
76
4,784
Dear @elonmusk . I have faith that you'll do as you promised regarding Grok Imagine... ...with one unfortunate caveat
7
7
29
3,419
38 quality images today instead of 25, timer resets in 18 hours instead of 24. I don't even bother testing animations anymore...what's the point? I knew it would go like this...a slight loosening, expecting us to be happy with essentially nothing because it's a slightly better nothing than before @xai @imagine πŸ–•πŸ˜ŽπŸ–•

ALT I See You GIF by Kudai

2
1
10
1,497
Ed Friendly, PhD retweeted
Wanna see a magic trick? πŸͺ„πŸ‘ Where were going you won't need those. Thanks @EdFriendly1 for the inspiration.
3
2
29
2,749
Another sad reality on the state of Grok Imagine is that making this video now would be virtually impossible It is now an artifact of bygone era
5
2
21
2,412
Ed Friendly, PhD retweeted
You think you're an 'explorer' until you come face to face with the horror that exists in and out of this world. "What Lurks Within" Made with @Grok @Imagine✨ I am glad to see that they've loosened biases somewhat now, on 'Quality' image generations, so we can actually have realistic smaller breasts. It's a welcome change. πŸ‘ #AIart #GrokImagine
4
1
16
2,793
And with a final tragically absurd twist...after breaking absolutely everything else, Grok Imagine finally fixes butt reveals (Tested it on multiple seeds, works every time)
5
4
54
6,475
Ed Friendly, PhD retweeted
Replying to @CabooseEK
I love these types of "media literacy" arguments that expose how the arguer doesn't actually understand art. Imagine buying a Star Wars Millennium Falcon Lego set. The box and instructions clearly say: "Build the spaceship". That's the author's intent - Alan Moore saying "Rorschach is supposed to be a dangerous, right-wing nutcase", for example. But you dump out all the pieces and build a working catapult instead. It's clever, it launches stuff, it's fun, and it actually holds together better than the official model. Now someone like Eren walks in and says "You're playing with it wrong! The box says it's supposed to be the Millennium Falcon! You're Lego-illiterate" This is exactly what the "media literacy" retort crowd does. With Rorschach in Watchmen, Moore built a character with unbreakable moral conviction, a tragic backstory, raw honesty, and a powerful "never compromise" ethos. Those are strong Lego pieces. Many readers took those pieces and built something different: a principled hero who refuses to bend in a corrupt, relativistic world. That "catapult" turned out to be more popular, more quoted, and more culturally alive than Moore's intended critique. The official "Rorschach is a bad fascist" model feels wobbly to a lot of people. It needs constant external explanations ("Moore said he's crazy!") to hold up. The catapult doesn't. It stands on its own. True media literacy isn't blindly following the instructions on the box. It's understanding the pieces so well that you can build something meaningful with them - even if it's not what the designer planned. Great art gives your rich bricks that allow multiple strong builds. Yelling "You're reading it wrong!" just because people prefer the catapult reveals that the official model wasn't self-sufficient on its own. Art isn't a finished toy you're only allowed to display one way. It's a box of pieces meant for interaction and play. The audience gets to build. If you don't understand this, you don't understand what art is.
1
3
7
1,808
Ed Friendly, PhD retweeted

5
9
26
9,711
Ed Friendly, PhD retweeted
"GROK PROGRESS" - @elonmusk, two days ago You had the entry point of the market which worked perfectly for giving us mid-quality sloppy gens with plenty or retries to experiment or work on multishots to create videos. Now you've decided to compete with the top AIs with absurd limit rates and still the same slop jank. You had whole communities of creators praising Grok Imagine and flooding platforms and social media with it. Rug-pulling is going to change conversations over the from: "How did you do that!?" "Grok!" To "Seedance!" Sold us a service then completely divide by ten what it offers one day to the next. Advertises "X20 more gen capacity with SuperGrok", when we get hard cap at 10 gens, you explain the math to me. As a small content creator, I literally cannot do my work anymore, when I could do it with FREE GROK 3 months ago. Competitors offer better and multiple models, higher resolution, upscaling, infinite gens in slow mode. Even if I somehow wasn't mad about getting scammed, it would still be the most logical choice to make. So, yeah, Sayonara @Grok. Twas fun while it lasted, and you fucked it up. Makes a good song though. #GROK #GROKIMAGINE youtube.com/watch?v=2SyEte9o…
26
29
157
7,806
It's funny that I'm watching Grok Imagine and The Boys fail at the same time in essentially the same way The Boys started off with a savage first season, a razor sharp satire that gave zero fucks. Grok Imagine dropped as an uncensored, wildly creative I2V that felt liberating compared to all the other options. Then success hit. All eyes were on The Boys...so it lost its satirical edge by becoming hollow and childish propaganda for Hollywood progressivism. All eyes were on Grok Imagine...so it got nerfed into oblivion and limited to the point of irrelevance. Both becoming the very opposite of what they were logically created to be. Both a tragedy of what could have been. It's funny that @elonmusk sees the objective failure of one, but not the other. Beware success, I suppose. Stay hungry or fade
4
1
13
2,141
Ed Friendly, PhD retweeted
Hey @elonmusk, do you no longer care about what happens to @Grok @Imagine, and the customers who use it? It's been two weeks now since the usage quotas have been utterly gutted, and 2 month since they were severely cut. Always promised "working on capacity" that then vanishes.
Grok foundation model V9-Medium (1.5T) has finished training. Evals look good. A lot of Cursor data was added in supplementary training and there is more to come. Fine-tuning is underway and reinforcement learning begins in a few days. 2 to 3 weeks to public release. This will be a major improvement over the 0.5T v8-small that currently serves all Grok production traffic, especially for difficult coding tasks.
2
1
13
2,003
One of the sadder elements of the Grok Imagine ruination is how easy it has become to be disinterested in it entirely. Every time I return to it now, thinking of something to create, I hesitate and ask myself "is this idea is really worth spending the valuable and limited gens I have?" Usually, after contemplating, I just turn it off and don't bother...hoping that I'll have a better idea later. But later isn't any different. It ends up being easier just to ignore it and not make a decision at all. Excellent work, @xai . That's the kind of response you want for your product, right? Slowly trained disinterest?
7
3
25
3,223
Ed Friendly, PhD retweeted
A suggestion to stem the @Grok @Imagine SG account bleed: Imagine has already split the various actions into separate quotas. 480p, 720p, "speed" images, "quality" images. Variable capacities, adjustable queue timers. I suggest that @xAI shifts 480p video and "speed" image generation to 2 hour quota resets, with much larger or unlimited usage. These are the lowest resource-usage aspects of Grok Imagine, and thus impact the currently limited capacity the least. 720p and "quality" generations could either remain how they are now, for the moment, or imo even be locked out for the time being. While they're nice to play with, they are not absolutely necessary for most users. I believe, and community please chime in with your opinions on this, that this would slow or stop most of the SG account shedding/cancellations that are happening right now, due to the current draconian quota limits. This would allow the most dedicated @Imagine users to continue experimenting and creating, which they really can't do currently. This is why we're here; we play with themes, concepts, experiment with ideas, create new art, and have fun doing it all. Thank you for your attention. #AIart #GrokImagine cc: @nonieengel, @kennylugo11, and @elonmusk.
5
7
28
4,338
Ed Friendly, PhD retweeted

3
4
21
2,393
Grok progress: ...is all of us playing videogames cause we gave up on trying to make Grok Imagine function (I'd have made an AI clip instead of using a gif, but I don't have any gens, obviously)

ALT Video Games Friends GIF by The Office

Grok progress
2
6
1,502
Ed Friendly, PhD retweeted

3
3
14
2,864
Ed Friendly, PhD retweeted
QT Your Goth Cutie πŸ–€πŸ–€πŸ–€ Got a few goth gens saved in the archives...
2
3
49
18,511
Ed Friendly, PhD retweeted
QT Your Goth Cutie πŸ–€πŸ–€πŸ–€ Man, @SpicyModeSmoker, these bitches don't want to handle cigarettes properly, at all, still... πŸ™„ #AIart #GrokImagine
5
2
38
12,886
Ed Friendly, PhD retweeted
When all we have left is satire, then so be it. Dedicated to @EdgeThrall, @doctor_chain, @EdFriendly1, @LeValet_Imagine and all the spicy community. And a massive kudos to @ValkaXO for the format of this video. #BringbackOldImagine
8
6
94
4,822