Jeremy has been writing about technology and PCs since the 90nm Netburst era (Google it!) and enjoys nothing more than a serious dissertation on the finer points of monitor input lag and overshoot followed by a forensic examination of advanced lithography. Or maybe he just likes machines that go “ping!” He also has a thing for tennis and cars.
You're worried about AI. We're worried about AI. But there's actually no need to worry. Because this is fine. For proof, all you need to do is ask generative AI to iterate on the"This is fine" cartoon meme. Pretty quickly, you'll agree that it is indeed all fine. Either that or you conclude this is completely bonkers, circular, meta, self-referential and beautifully deranged. And you'll be weeping with laughter like me.
The ruse goes like this. Ask a text-based AI model like CoPilot or ChatGPT to feed a prompt for"a comic of a dog saying 'this is fine'" into an image generator such as DALL-E 3 and the output will likely be pretty much exactly the same as the original cartoon. This is indeed where my favourite new AI rabbit hole has come from: aThat original, for the record,Anyway, it's when you ask the model to iterate on that initial output where the madness and hilarity ensues.
The utter insanity into which the images rapidly descends is, all at the same time, totally hilarious and faintly alarming, but also perfectly encapsulating of the current state of AI.This is AI in its purest form. Totally mad, utterly alien and yet disarmingly human, chaotically creative, subtly knowing and somehow destructively oblivious.
Sure, you'll get some slightly alarmingly deformed cartoon dogs. But you'll also get gibberish text that sort of makes sense at the same time as hinting at a nascent consciousness that almost certainly isn't there but which our human instincts can't help but perceive. And weird visual-to-text mashups where the delineation between text characters and images are blurred. And just pure, giddy, thrilling, hilarious madness.
日本 最新ニュース, 日本 見出し
Similar News:他のニュース ソースから収集した、これに似たニュース記事を読むこともできます。
Copilot can't stop emitting violent, sexual images, says Microsoft whistleblowerAI tech loves picturing women in underwear, Windows giant won't do anything about it, watchdog told
続きを読む »
Microsoft trying to stop Copilot generating fake Putin comments on Navalny's deathPlus: US Department of Justice has hired its first chief AI officer
続きを読む »
Microsoft Introduces Copilot for Security, a Generative AI ServiceMicrosoft has launched Copilot for Security, a generative AI service that offers document summarization, flagging suspicious interactions, and generating security recommendations. It can be used as a standalone portal or integrated with third-party products. The service is based on GPT-4 and a Microsoft security-specific model and is available through a pay-as-you-go licensing model tied to Microsoft Azure.
続きを読む »
Microsoft forges One Teams App To Rule Them AllWindows preview delivers one app to handle multiple accounts, boosts Copilot's capabilities
続きを読む »
ChatGPT-maker OpenAI hits back at Elon Musk criticismOpenAI rejects claims it is focusing too much on profit, saying it is an idea Elon Musk endorsed.
続きを読む »
Elon Musk sues ChatGPT-maker OpenAI over Microsoft linksThe lawsuit accuses the firm of abandoning its founding principle of benefiting humanity.
続きを読む »