But does it actually scan? QR isn’t actually very hard to make, but 0% chance imagegen can make one encoded properly, so that would tell us it’s calling useful functions correctly (bet it ain’t).
November 9, 2026 at 8:04 PM
But does it actually scan? QR isn’t actually very hard to make, but 0% chance imagegen can make one encoded properly, so that would tell us it’s calling useful functions correctly (bet it ain’t).
It is significantly to Anthropic's credit that they do not and apparently as a point of principle will not do imagegen
November 2, 2026 at 2:34 PM
It is significantly to Anthropic's credit that they do not and apparently as a point of principle will not do imagegen
yeah and it's also hard to teach that a TRUE is absolute but a FALSE is just a "maybe not"
makes me feel fucking prescient back in 2018 when I told people the main problem with imagegen is gonna be the uncertainty and not the fakes
makes me feel fucking prescient back in 2018 when I told people the main problem with imagegen is gonna be the uncertainty and not the fakes
November 4, 2026 at 10:22 PM
yeah and it's also hard to teach that a TRUE is absolute but a FALSE is just a "maybe not"
makes me feel fucking prescient back in 2018 when I told people the main problem with imagegen is gonna be the uncertainty and not the fakes
makes me feel fucking prescient back in 2018 when I told people the main problem with imagegen is gonna be the uncertainty and not the fakes
these are so... quaint. Really "off", like modern imagegen by default, but chill. It's nice. I particularly like the stone... snake... lady-thing?
November 3, 2026 at 11:43 PM
these are so... quaint. Really "off", like modern imagegen by default, but chill. It's nice. I particularly like the stone... snake... lady-thing?
til: asked the claude app to get help from codex's imagegen. it did. and it worked.
November 4, 2026 at 10:11 PM
til: asked the claude app to get help from codex's imagegen. it did. and it worked.
yeah agree, it would be nice to at least know the limits. though, like, it wouldn't be *that* hard to test empirically - generate something with openai imagegen, then transform and submit until the system stops detecting it.
November 4, 2026 at 10:19 PM
yeah agree, it would be nice to at least know the limits. though, like, it wouldn't be *that* hard to test empirically - generate something with openai imagegen, then transform and submit until the system stops detecting it.
it's the standard way chatgpt and codex create images. in my case it was my agent using imagegen from codex.
November 1, 2026 at 1:33 PM
it's the standard way chatgpt and codex create images. in my case it was my agent using imagegen from codex.
it's a little complicated. the short answer is that i don't know 100% that it's AI, but this is an execution of a well-known airbrush style and it's incorporating a few highlight quirks that you often see with imagegen stuff even when it's ripping a particular style. that and it's coming from chuds
November 3, 2026 at 6:11 PM
it's a little complicated. the short answer is that i don't know 100% that it's AI, but this is an execution of a well-known airbrush style and it's incorporating a few highlight quirks that you often see with imagegen stuff even when it's ripping a particular style. that and it's coming from chuds
it's pretty fun to work on 3d models like this
unlike imagegen-based 3d construction approaches which puke out triangle soup, it actually (sort of) understands what it is doing, and gives parts/objects semantic names, and you can refer to those to adjust them
unlike imagegen-based 3d construction approaches which puke out triangle soup, it actually (sort of) understands what it is doing, and gives parts/objects semantic names, and you can refer to those to adjust them
November 1, 2026 at 12:49 AM
it's pretty fun to work on 3d models like this
unlike imagegen-based 3d construction approaches which puke out triangle soup, it actually (sort of) understands what it is doing, and gives parts/objects semantic names, and you can refer to those to adjust them
unlike imagegen-based 3d construction approaches which puke out triangle soup, it actually (sort of) understands what it is doing, and gives parts/objects semantic names, and you can refer to those to adjust them
But but characterai exploded the dominoes are falling now for realsies
(i still want imagegen to get nuked tho tbh)
(i still want imagegen to get nuked tho tbh)
October 30, 2026 at 5:52 AM
But but characterai exploded the dominoes are falling now for realsies
(i still want imagegen to get nuked tho tbh)
(i still want imagegen to get nuked tho tbh)
I think it's less that and they simply don't want any part of it and it's not their focus (they are very much of the "we want to be positive AI!" mindset and there's very little positive about imagegen)
November 2, 2026 at 5:58 PM
I think it's less that and they simply don't want any part of it and it's not their focus (they are very much of the "we want to be positive AI!" mindset and there's very little positive about imagegen)
Turn a portrait into a hyperrealistic handcrafted puppet in the style of the Pedro monkey meme.
gbti.network/prompts/pedr...
#imageprompt #prompthack #fun #portrait #Pedro #sidestare #muppets #textures #gpt #imagegen #entertainment #kento #gbti
gbti.network/prompts/pedr...
#imageprompt #prompthack #fun #portrait #Pedro #sidestare #muppets #textures #gpt #imagegen #entertainment #kento #gbti
Pedro Monkey Puppet Portrait | GBTI Network
A photograph of the subject rebuilt as a felted puppet: densely flocked orange skin with individually resolved fibres, glossy protruding eyes behind their own glasses, their real hairstyle and beard r...
gbti.network
November 2, 2026 at 10:30 PM
Turn a portrait into a hyperrealistic handcrafted puppet in the style of the Pedro monkey meme.
gbti.network/prompts/pedr...
#imageprompt #prompthack #fun #portrait #Pedro #sidestare #muppets #textures #gpt #imagegen #entertainment #kento #gbti
gbti.network/prompts/pedr...
#imageprompt #prompthack #fun #portrait #Pedro #sidestare #muppets #textures #gpt #imagegen #entertainment #kento #gbti
Day 30: Giving exceptions a place.
After expanding the reports, the build added a data-quality panel: a summary of monitored forms and a closer look at flagged entries. It was a move toward bringing the follow-up work into the portal, alongside the information staff needed.
After expanding the reports, the build added a data-quality panel: a summary of monitored forms and a closer look at flagged entries. It was a move toward bringing the follow-up work into the portal, alongside the information staff needed.
November 2, 2026 at 1:05 PM
Day 30: Giving exceptions a place.
After expanding the reports, the build added a data-quality panel: a summary of monitored forms and a closer look at flagged entries. It was a move toward bringing the follow-up work into the portal, alongside the information staff needed.
After expanding the reports, the build added a data-quality panel: a summary of monitored forms and a closer look at flagged entries. It was a move toward bringing the follow-up work into the portal, alongside the information staff needed.
But but characterai exploded the dominoes are falling now for realsies
(i still want imagegen to get nuked tho tbh)
(i still want imagegen to get nuked tho tbh)
October 30, 2026 at 5:52 AM
But but characterai exploded the dominoes are falling now for realsies
(i still want imagegen to get nuked tho tbh)
(i still want imagegen to get nuked tho tbh)
夢つづりアニメさん、こんばんは😀
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
October 31, 2026 at 11:32 AM
夢つづりアニメさん、こんばんは😀
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
Tired of your image generator sanitizing every prompt? Stop settling for "policy-safe" results. We name the 2026 models that actually bypass restrictions on nudity and violence to deliver what you actually asked for. #AI #imagegen
https://aethersky.ai/learn/uncensored-ai-image-generator-2026
https://aethersky.ai/learn/uncensored-ai-image-generator-2026
October 30, 2026 at 8:00 PM
Tired of your image generator sanitizing every prompt? Stop settling for "policy-safe" results. We name the 2026 models that actually bypass restrictions on nudity and violence to deliver what you actually asked for. #AI #imagegen
https://aethersky.ai/learn/uncensored-ai-image-generator-2026
https://aethersky.ai/learn/uncensored-ai-image-generator-2026
夢つづりアニメさん、こんばんは😀
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
October 31, 2026 at 11:32 AM
夢つづりアニメさん、こんばんは😀
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
嬉しいレスをありがとうございます🤗
プロンプトはGemini、Grok、Canva、ChatGPTで利用できますが、絵の揺れはChatGPTでしか比較していません。残念ながら、このように節に分けても揺れます。節に分ける目的は、素材ページに記したように、構造化して生産性を上げる（2次利用しやすいようにする）ためで、それは達成できています。
CodexのImagegenスキルを使うと、自己チェックが働いて揺れが減りますが、これはAIエージェントのメモリが効いているためです。
Make your best CV result easier to spot: put the most relevant result in that role’s first bullet.
7.4s: Ladders’ 2018 initial-screen average. Colour highlights document hierarchy.
Save for your next CV edit.
#CVTips #ResumeTips
7.4s: Ladders’ 2018 initial-screen average. Colour highlights document hierarchy.
Save for your next CV edit.
#CVTips #ResumeTips
October 30, 2026 at 8:00 AM
Make your best CV result easier to spot: put the most relevant result in that role’s first bullet.
7.4s: Ladders’ 2018 initial-screen average. Colour highlights document hierarchy.
Save for your next CV edit.
#CVTips #ResumeTips
7.4s: Ladders’ 2018 initial-screen average. Colour highlights document hierarchy.
Save for your next CV edit.
#CVTips #ResumeTips
i'm so with u. imagegen was way more interesting when it spit out fucked up looking stuff
October 27, 2026 at 7:45 PM
i'm so with u. imagegen was way more interesting when it spit out fucked up looking stuff
MAI-Image-2.6 and MAI-Image-2.6-Flash are now generally available in Microsoft Foundry
buff.ly/GWMfOTz
#foundry #ai #azure #mai #models #microsoftfoundry #cloud #imagegen
buff.ly/GWMfOTz
#foundry #ai #azure #mai #models #microsoftfoundry #cloud #imagegen
MAI-Image-2.6 and MAI-Image-2.6-Flash are now generally available in Microsoft Foundry | Microsoft Community Hub
From creative exploration to enterprise workflows Creating a compelling image is one step. Making image generation useful across a business means giving...
buff.ly
October 28, 2026 at 1:00 PM
MAI-Image-2.6 and MAI-Image-2.6-Flash are now generally available in Microsoft Foundry
buff.ly/GWMfOTz
#foundry #ai #azure #mai #models #microsoftfoundry #cloud #imagegen
buff.ly/GWMfOTz
#foundry #ai #azure #mai #models #microsoftfoundry #cloud #imagegen
Pair keys with action labels so players know what each input does.
Quiet Keys: 12 keys, normal/pressed, free SVG/PNG. AI-assisted design/code; no imagegen.
willlive.itch.io/quiet-keys-f...
#GameAssets #GameDev #dots
Quiet Keys: 12 keys, normal/pressed, free SVG/PNG. AI-assisted design/code; no imagegen.
willlive.itch.io/quiet-keys-f...
#GameAssets #GameDev #dots
October 26, 2026 at 12:11 AM
Pair keys with action labels so players know what each input does.
Quiet Keys: 12 keys, normal/pressed, free SVG/PNG. AI-assisted design/code; no imagegen.
willlive.itch.io/quiet-keys-f...
#GameAssets #GameDev #dots
Quiet Keys: 12 keys, normal/pressed, free SVG/PNG. AI-assisted design/code; no imagegen.
willlive.itch.io/quiet-keys-f...
#GameAssets #GameDev #dots
Possibly the largest crime enabled by GenAI
Horrible all hands meeting slides with bad imagegen
Horrible all hands meeting slides with bad imagegen
October 22, 2026 at 4:20 PM
Possibly the largest crime enabled by GenAI
Horrible all hands meeting slides with bad imagegen
Horrible all hands meeting slides with bad imagegen
i got bored of imagegen/video/sunoslop and figured: what if demoscene?
here's ring0, a 64k intro for native macOS arm64, single self-contained 53,304 byte Mach-O, no static assets used, all completely synthesized in C and arm64 assembly.
opus 5.5, "make a native 64k intro" and a single moodboard
here's ring0, a 64k intro for native macOS arm64, single self-contained 53,304 byte Mach-O, no static assets used, all completely synthesized in C and arm64 assembly.
opus 5.5, "make a native 64k intro" and a single moodboard
October 22, 2026 at 2:23 PM
i got bored of imagegen/video/sunoslop and figured: what if demoscene?
here's ring0, a 64k intro for native macOS arm64, single self-contained 53,304 byte Mach-O, no static assets used, all completely synthesized in C and arm64 assembly.
opus 5.5, "make a native 64k intro" and a single moodboard
here's ring0, a 64k intro for native macOS arm64, single self-contained 53,304 byte Mach-O, no static assets used, all completely synthesized in C and arm64 assembly.
opus 5.5, "make a native 64k intro" and a single moodboard
It was a shocker to see colleagues I recognize’s likenesses being used in ImageGen posters
October 22, 2026 at 4:22 PM
It was a shocker to see colleagues I recognize’s likenesses being used in ImageGen posters