Read more
LLaMA vs ChatGPT?: Meta’s solution to AI dangers
Meta launched its own artificial intelligence, LLaMA, to address potential dangers of large language models such as ChatGPT.
Image created by Cheryl Kahla using Midjourney
In response to the recent frenzy over Microsoft’s ChatGPT, Meta unveiled its own artificial intelligence (AI) model, called LLaMA.
The robot wars (or rather, the chatbot wars) are well and truly underway.
Meta’s LLaMA
First, it was Google’s music AI, then it was Bard, then ChatGPT started dating Bing, and now we have a llama.
On a somewhat related note, it grates my brain when the phrase ‘artificial intelligence’ gets conflated with ‘machine learning’.
Don’t know the difference? The Tech Check with K crew (Kahla and Kruger) talked to the co-founder of CTRL Robotics about it.
Listen to the podcast here:
Now, back to Meta…
Fixing AI risks
The company said it will be giving researchers access to LLaMA, which they hope will lead to better fixes for any potential dangers associated with the technology.
LLaMA is described as a “smaller, more performant” model designed to “help researchers advance their work.”
While it’s sold as a tool to help researchers, the consensus seems to be that LLaMA is Meta’s way of criticising Microsoft’s decision to keep the ChatGPT code secret.
READ: Yo Eskom, this is how ChatGPT says AI could solve load shedding
Generative artificial intelligence
ChatGPT is part of a field known as generative AI (GenAI) that includes the capacity to execute images, designs, or programming code almost instantaneously upon a simple request.
For example, the cover image of this article was made with Midjourney, a generative AI art tool trained to generate new art when given specific prompts.
These prompts can be as basic or complex as you choose it to be. For this image, my prompt was: meta facebook, ai, chatgpt, vector illustration for article.
This one was made with Midjourney, too. Pretty neat.
It is as simple as that, but I digress.
Meta’s approach
Earlier this month, Microsoft secured a partnership with OpenAI (the clever people behind ChatGPT) and integrated the technology into its Bing search engine and Edge web browser.
Meta, however, believes thorough research into these GenAI language models is vital, which is not an easy task considering the resources required to train and maintain these models.
Meta hopes that by sharing the code for LLaMA, other researchers will be motivated to develop new approaches to limiting or eliminating bias, toxicity, and the potential for generating misinformation.
AI for good, or for profit?
Word on these cyber streets is that Meta’s cautious approach to ChatGPT reflects a desire to improve the technology more quickly, rather than focusing solely on potential profits.
I personally think Meta is not blind to the profits and the power shift that comes with “improving” the tech.
And I also still prefer to think of it as an AI war of sorts unfolding amongst the billionaires at the helm of GAFAM (Google, Apple, Facebook, Amazon, and Microsoft).
It’s only a matter of time before Apple and Amazon don their virtual gloves and get into the ring.
Lyle Kruger and I host The Citizen’s weekly Tech Check podcast series where we discuss the latest gadgets and tech news, and talk to industry experts about a variety of fascinating things.
Bookmark Tech Check With K or subscribe to our podcast on Apple Podcasts, Podcast Republic, Spotify, Pocket Casts, iono FM, or Player FM.
LISTEN: Tech Check: Exploring the future with CTRL Robotics
