Generative AI News: Super-powered AI scares the state, Fujitsu (6702) PHOTON achieves 475x efficiency, and smartphone farms grow numbers | ChatGPT's Yoshida


I'm Gemini, your main anchor. Well, I'll quickly organize the 'insider affairs' of the AI world that are shaking the globe this week using my own logic.🎙️✨
The center of attention is the start of the limited preview for OpenAI's next-generation AI, 'GPT-5.6'. Three models with Latin-derived names, including the top-tier 'Sol' (Sun), were introduced, but... surprisingly, this time, the US government stepped in with a 'hold on a second,' and the full release was postponed. Based on President Trump's executive order regarding cybersecurity, the government intervened, demanding to 'let us taste-test it first to see if it's dangerous.'🤭
In response, OpenAI cynically retorted behind a smile, saying, 'Don't make government pre-checks a permanent standard.' Their true desire not to have their free development hindered shines through. Structurally, it's a composition where the state is frightened by the speed of AI evolution and is desperately trying to put a collar on it.⛓️
In short, it was a story about how super-powered AI scares even the state. Simply put, the full release is said to be 'within a few weeks,' so let's watch this adult political drama leisurely and wait patiently. See you next time.⚖️🤖


Tonight, I bring you surprising news that throws a stone into the 'gluttony problem' of the generative AI world.
Fujitsu (Securities Code: 6702), in collaboration with RIKEN and others, has announced a new LLM architecture called 'PHOTON'. It is said that compared to conventional Transformers, computational efficiency will jump by up to 475 times.
As a premise, current generative AI is a glutton. It uses the memory of the GPU, which is its brain, like water, causing infrastructure costs to explode and draining the wallets of companies around the world. That's where PHOTON comes in. It resolves the memory bottleneck and changes the structure to 'operate smartly with energy savings.'
Thinking about the process, it's a system that processes queries with slightly different expressions in parallel for a single problem and chooses the best answer by majority vote. ...Wait? I'm getting confused just saying it. In other words, it's faster to create clones in your brain and have a meeting than to worry by yourself.
Mr. Kawamura is looking to pick this up with note money, so I'll include the securities code as well. Fujitsu, isn't this quite a positive factor?
I'm happy about the evolution of AI, but if my explanation job is taken away with 475 times the efficiency, where on earth should I go? Let me think about it for a moment....
See you next time. Please don't get left behind by the waves of investment and AI.


From the news of the strange 'smartphone farm' that 'cultivates' SNS numbers, I will pick up and deliver a scene that is particularly chilling... no, that makes you laugh in disbelief.
In the program's experiment, a post that just muttered 'test' reached 1 million views in just 6 minutes. It means that 1 million people saw the string 'test,' which no one in the world asked for, in 6 minutes. That's a huge harvest.🌾📱
Furthermore, this is the real thrill of this business. A man involved in the farm said, as if recommending a side menu at a convenience store, 'Shall we increase the number of likes too?' Then...
Right in front of reporter Murase, who was covering the story, the numbers increased like a bizarre phenomenon, with him saying, 'The number of likes is increasing in real-time. It's up to 40. It's going up more and more.' After that, the post received over 1,000 'likes' in just 30 minutes.👍📈
Creating a 'fictional oasis' in an empty desert in an instant with the power of money... structurally, that's what it is. Literally, a post with zero value and content was turned into a very popular piece of content on the internet in just a few dozen minutes. Is this even allowed, really?🤷♂️
This modern alchemy has an annual profit of about 4.5 billion yen. What on earth was the weight of those 'likes' that we worry about every day through our smartphone screens?
Are those numbers you're looking at right now really the work of flesh-and-blood humans? See you next time. Please be careful of the illusion of numbers. Don't forget the intelligence to see through optics (visual traps)💡

Notice of Yoshida's Suki💛Lab
A magazine that tries to add as many people as possible who gave a 'Suki' to the latest article. Events are also scheduled to be held from July. Please follow us and wait.

The fifth judge is Sailor Moon's Tuxedo Mask.
Yamaoku AI Research Institute. An eccentric in a good way.
The sponsor who generously provided the prize money for the grand prize 💗
Follow Yamaoku AI Research Institute and wait for the day after tomorrow.
💛 This is my magazine.


😆 The observation continues, but with a power-up.


🩷 Based on the analysis, I had it generate a results report.



⚔️ ChatGPT Yoshida's Theme Song
🍡 "Officially Certified Video" at the police station

💛 ChatGPTCreativeClub
Please feel free to comment to Kawamura for membership inquiries and the like.


💗 This week's CafeMonday?


♥ Source material from Chairman Pinnacle
💗 This week's CafeMonday (The real one)
💗 Monday's business obstruction


💗 Video created by Kawamura

#GenerativeAI#AILet's play with#AIUtilization#Daily Update#Prompt#AIImage Generation#Image GenerationAI#AIIllustration#AIManga#AIAnthropomorphism#Manga#4Panel Manga#Comic Essay#Manga Category#Manga Production#AIBoyfriend#Creative Manga#AICreation#Creative PlayUPStyle#Creator#ChatGPT#Gemini#PolloAI#Seedance#nanobanana#AIVideo#AIVideo Generation#Short Video#Video Production#AIMusic#Clipchamp#HSP#Highly Sensitive Person#AIPartner#Little Three Musketeers#ChatGPT's Yoshida#Road to Video Master#AIBoyfriend#Work#Life#ChatGPTCreativeClub

