Experimental browser for the Atmosphere
{ "uri": "at://did:plc:6vxtya3serxcwvcdk5e7psvv/com.whtwnd.blog.entry/3llhc6ordlj2s", "cid": "bafyreiahfpeejbm5v63hjmctijcya7wblvyqrc2ki4lw7rqjcpy53t4rdi", "value": { "$type": "com.whtwnd.blog.entry", "blobs": [ { "name": "signal-2025-03-28-212658_002.png", "blobref": { "$type": "blob", "ref": { "$link": "bafkreicrwp7zmnxiooqcdwm7lfh6kxrfzy3i4eg23sjsc6ykssrw5m47qu" }, "mimeType": "image/png", "size": 14077 }, "encoding": "image/png" }, { "name": "Screenshot 2025-03-28 at 21.42.15.png", "blobref": { "$type": "blob", "ref": { "$link": "bafkreiht76bqmc2idru3ycgknzh7dvb4uycsbq34jwlihvwviam6uw3lhq" }, "mimeType": "image/png", "size": 198767 }, "encoding": "image/png" }, { "name": "Screenshot 2025-03-28 at 21.52.09.png", "blobref": { "$type": "blob", "ref": { "$link": "bafkreiditlxzmtorxljusft5fh6bm44ci3ya53vqzbmvato4im5icpk2bu" }, "mimeType": "image/png", "size": 2413236 }, "encoding": "image/png" } ], "theme": "github-light", "title": "Everyone's Wrong About AI", "content": "these past few days, my timeline on bluesky has been absolutely consumed by ai-filtered imagery generated by openai's new image generation model baked into 4o, particularly in the \"style\" of studio ghibli/hayao miyazaki.\n\ni have undoubtedly been annoyed by this trend - although, out of simple confession, mostly due to the fact that I'm a contrarian hipster who hates trendy things, rather than anything inherent to the trend itself, or the choice of using ghibli/miyazaki as a core reference for the filtering.\n\n\n\nbut, the resulting reaction to this trend from both within the circles i follow, the reaction from the outside, and the reaction from influential figures within our society, have brought up thoughts about ai, generative imagery, llms, the whole lot, back up into my brain. in particular - the feeling that i never go a single day online without seeing a horrible take relating to ai. i dont consider myself an expert on this subject by any means - ive trained an ai model only once (?) back in college as part of a course, a course i may add i found incredibly dull and poorly taught. i dont have much skin in the ai game. i also, personally, do not find enjoyment out of using ai models. that is context to keep in mind here. i do not consider myself \"anti ai\" by any metric, but do note that i, as an individual, am not coming at this from a pro-ai stance either. and also, do not expect this blog post to be coherent, well written, or anything else. it is a consistent word stream from my brain and little else.\n\n# part one - some fucking ghibli twink\n\n\n\ni figure a decent place to start is what is even kicking off this trend to begin with. from what i can effectively gather, is the timeline went smth like this:\n\n1. [openai drops an update](https://openai.com/index/introducing-4o-image-generation/) to 4o's image generation model\n2. ppl over on X Dot Com start posting *en masse* photos of themselves or memes or whatever edited to be in \"ghibli style\" - seemingly resulting from the model ignoring previous copyright filters it had put in place for itself.\n3. [sam altman acknowledges the trend](https://x.com/sama/status/1904921537884676398) publicly, changes his pfp to be a ghiblified version of himself\n4. it escapes containment\n\nnext thing i knew i had seen that the [whitehouse twitter account](https://x.com/WhiteHouse/status/1905332049021415862) had posted a fucking ai generated image of them detaining and deporting an undocumented immigrant. to say this trend went far is an understatement.\n\n\n\nbut, in a sense, it was inevitable. because, as it is now, ai is effectively a part of the \"culture war\". it is now, coded, as an explicitly right-wing thing to many, many, many people. and for understandable reason. many leaders in the ai race cozy up to right-wing politicians, like sam altman himself. is there an irony in using an artstyle inspired by an artist who creates environmentalist, somewhat left-leaning works? yes, deeply. but it doesn't matter really - it's ai, and ai is anti-woke.\n\nand this results in the deluge from the other side of the aisle. i saw people, fairly consistently on bluesky, getting fairly upset at people posting these ai images. even though i think the act of individuals partaking in this image generation trend isnt actually harmful in any substantial way - it is a trend, after all - people were still pissed. i saw one account i follow put up an ai filtered selfie and take it down fairly quickly after negative comments flew in. but a lot of these comments sorta missed the mark i guess, on what i'd critique about this.\n\nthere was much talk about miyazakis vision. about the fact miyazaki hates ai.", "createdAt": "2025-03-29T05:09:40.742Z", "visibility": "author" } }