From the company that created ChatGPT, things get hotter than ever.
The AI field was dull and quiet, and the buzz it created mostly happened within its own field, and rarely reach far beyond its own audience. But when OpenAI introduced ChatGPT as a AI chatbot tool, the internet was quickly captivated.
Powered by GPT-3.5, the AI is able to do a wide range of tasks, including writing poetry, technical papers, novels, and essays.
And when others, like Microsoft, started embedding the technology into Bing and Edge, found issues that the AI can get emotional at times, and then fixing it, the hype continues to grow.
This time, after building on the hype, and even monetize from it, OpenAI is ramping things up into yet another step, by introducing GPT-4.
According to OpenAI in its web page:
" It can generate, edit, and iterate with users on creative and technical writing tasks, such as composing songs, writing screenplays, or learning a user’s writing style."
Announcing GPT-4, a large multimodal model, with our best-ever results on capabilities and alignment: https://t.co/TwLFssyALF pic.twitter.com/lYWwPjZbSg
— OpenAI (@OpenAI) March 14, 2023
According to OpenAI, the long-awaited update of its AI technology is also making ChatGPT "safer and more aligned" than its predecessor.
The effort of scaling up this large language model through deep learning, allows it to exhibit "human-level performance" on some professional and academic tasks.
GPT-4 is also "more creative and collaborative than ever before."
In the end, the update should benefit other technologies that depend on it, allowing more output reliability.
For example, users using GPT-4 can send it a photo of the inside of a refrigerator, and GPT-4 will not only correctly identify what is in there, but also come up with ideas about what to cook with those ingredients.
And of course, ChatGPT will benefit from this update.
Here, OpenAI said that ChatGPT powered by GPT-4 is able to perform better than when it uses GPT-3.5, and should be less likely to "go off the guardrails" with its responses.
OpenAI said the updated technology passed a simulated law school bar exam with a score around the top 10% of test takers.
In comparison, GPT-3.5 scored around the bottom 10%.
GPT-4 can also read, analyze or generate up to 25,000 words of text, and write code in all major programming languages, according to the company.
Summarizing everything, GPT-4 can see and understand images, thanks to OpenAI's partnership with Be My Eyes. Its also harder to trick, and more difficult to derail. And because GPT-4 has a larger memory to begin with, the AI can remember longer conversations, spanning to about 64,000 words.
GPT-4 is more multilingual
It's worth noting, that even though GPT-4 was trained on vast amounts of online data to generate compelling responses to user prompts, and that OpenAI described the update as the "latest milestone" for the company, the company admits that it is still "less capable" than humans in many real-world scenario.
Because of this, the company said that “great care should be taken when using language model outputs, particularly in high-stakes contexts.”
"It is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it," said Sam Altman, CEO of OpenAI, in a series of tweets, suggesting that GPT-4 has similar limitations as earlier GPT models.
But there are noticeable improvements, he said.
here is GPT-4, our most capable and aligned model yet. it is available today in our API (with a waitlist) and in ChatGPT+.https://t.co/2ZFC36xqAJ
it is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it.— Sam Altman (@sama) March 14, 2023
"It is more creative than previous models, it hallucinates significantly less, and it is less biased," he said.
"We spent six months making GPT-4 safer and more aligned. GPT-4 is 82 percent less likely to respond to requests for disallowed content and 40 percent more likely to produce factual responses," OpenAI said.
Microsoft, as the biggest commercial backer of OpenAI, is the first third-party product to use GPT-4.
"We are happy to confirm that the new Bing is running on GPT-4, which we’ve customized for search," Microsoft said in a blog post. "If you’ve used the new Bing preview at any time in the last five weeks, you’ve already experienced an early version of this powerful model."
The news comes weeks after OpenAI announced it is opening up access to its ChatGPT tool to third-party businesses through APIs, paving the way for the chatbot to be integrated into numerous apps and services.
Further reading: Transitioning To AGI Is Perhaps The Most Important, And 'Scary' Project In Human History