- Smoking Robot
- 🤖 Microsoft Exec Says ChatGPT-4 Coming Next Week
🤖 Microsoft Exec Says ChatGPT-4 Coming Next Week
Updated ChatGPT, chatbot clones, mind-bending video, Headlines, Links and more
Hello. ChatGPT-4 sounds like it’s coming next week. Will the world change again?
In the email today:
ChatGPT-4 may be released next week 🤖
ChatGPT clones are many 👯♀️
World-bending AI video of the day 📼
Here we go.
Microsoft CTO Says ChatGPT-4 Set for Release Next Week❗️
The rise of ChatGPT continues unabated. Interest in and thirst for the technology seems to be everywhere.
Microsoft knows this, and is pressing its existent advantage over competitors, like Meta and Google, with a new and improved ChatGPT that enhances technology which many users haven’t even maximized yet.
Putting video features into ChatGPT is really strapping a rocket to it, which it seems is exactly what Microsoft is hoping for here.
Here’s what Microsoft’s CTO of Germany said about it:
Microsoft is set to launch GPT-4 as early as next week.
GPT-4 will offer multimodal models that will allow the creation of videos from text prompts as well as translating text into images and music.
ChatGPT currently only replies in text form, so this update which purportedly includes video capabilities is a serious step forward.
Facebook owner Meta launched Make-A-Video in 2022, which creates realistic videos based on short text prompts— GPT-4 projects to be a significant improvement on that.
Microsoft did not touch on its integration of ChatGPT into its Bing web browser at a recent AI event.
We don’t think Andreas Braun (Chief Technology Officer at Microsoft Germany) just carelessly let this news slip at an AI conference. It’s not like Braun was a few cocktails into an evening and just started talking.
Braun and by extension Microsoft knew that attaching the words “multimodal” and “video” to ChatGPT’s next iteration would excite people.
Until we get our hands on it, we are left speculating on the sorts of things users will be able to do with GPT-4. But there are some obvious probable use cases.
Creation of audiovisual presentations (advertisements, TikToks, Instagram posts, etc.) that presently require time and effort to put together will likely become the quick and dirty domain of GPT-4.
Less sexy was the example Microsoft gave of call centers being able to have GPT-4 summarize conversations between employees and customers in real time rather than the employees having to prepare the summaries after the call ends. Not sexy, but definitely a time-saver, and time is money.
Already, though, there are questions and concerns about what sort of trouble GPT-4 might get itself into given the many “hallucinations” and other glitches ChatGPT has put out in its brief time on Earth.
It’s one thing to invent an answer to a question that turns out to be inaccurate. It’s quite another to invent audiovisual content that probably cannot be fact-checked.
So yeah…we can’t wait to see what GPT-4 can do.
Meanwhile, ChatGPT Clones Are Already in the Works 🖥️
While Microsoft is giving ChatGPT even greater capabilities, many other companies are happy enough for the moment to create their own versions of the program.
Some of these “clones” of ChatGPT project to be open use. They may not be as good as the original, but if they cost less than what Microsoft is almost certainly going to charge, many users probably won’t care.
Competing projects and open-source clones of ChatGPT are being developed by startups like Stability AI, as well as companies like Anthropic, Cohere, and AI21.
The wider availability of ChatGPT-style bots may make the technology more accessible to developers and researchers, and accelerate the rush to make money with AI tools.
However, wider availability of the technology may also complicate efforts to predict and mitigate risks associated with it, such as making up facts, adopting problematic personas, and facilitating malicious tasks like producing malware code or disinformation campaigns.
Some researchers have called for deployment of ChatGPT-like systems to be slowed while risks are assessed, but the competition between companies suggests little appetite for slowing down and instead incentivizes proliferation of the technology.
Efforts to replicate ChatGPT were an inevitable outcome given its meteoric rise and popularity. Every preeminent consumer product has less expensive knockoffs in the market.
Hopefully, some of these ChatGPT clones will be effective enough to make the technology widely available such that it won’t only be in the hands of corporations and governments.
As for what bad actors might do with them, well, that was also an inevitable outcome.
World-Bending Demo Video of the Day 📼
What happens when ChatGPT gets access to everything that’s ever been on your screen?
@florian_jue & @its@itsalfredwided to see for themselves - this is a glimpse into the near future of omnipresent personal AI.
Powered by @Rew@RewindAIhttp
— Linus (●ᴗ●) (@LinusEkenstam)
Mar 9, 2023
You may have to click the link to see the video, but it features the alarmingly-named Limitless, an AI tool that tracks… everything that happens on your computer screen… and allows you to rewind and ask questions about anything that’s come across your screen.
Very cool, but also the sort of tool in SmokeBot’s view that will create an obscene amount of privacy concerns, revealing the very many weird habits people have it. Would recommend that teenage men don’t download this tool.
Microsoft announces that ChatGPT is now part of its Azure OpenAI Service: Presently there are around 1,000 customers using the platform. Billing for ChatGPT service within the platform begins on March 13. 💰️
United States Chamber of Commerce wants regulation of AI: The administrative agency is the latest governmental entity to be concerned that AI could “hurt growth or become a national security risk.” ⚖️
Meta concedes that it was too slow to deploy graphics processing units to create AI infrastructure: “We were a little bit late to the trends,” said Javier Olivine, Meta’s COO. “So we’re playing a little bit of catch-up.” 👀
AI improves accuracy of severe weather forecasting: Weather researchers at Colorado State University have developed a computer model that applies data from “past weather events from historical archives” to sharpen predictions of severe weather events. 🌪️
Unsurprisingly, the people most likely to make a lot of money on AI are the people who already had a lot of money to begin with 🤑
The communications industry is at or near the top of the list of businesses about to be “reinvented” by AI 👨💻
Famed linguist and philosopher Noam Chomsky calls chatbots “marvels of machine learning” but says they will never have the powers of the human brain 🧠
AI’s primary job loss targets (writers, artists, even lawyers) have a new pejorative moniker: “laptop class workers” 💻️