Welcome back to another episode of Tech Brood. This week we'll have some more tech news for you. We'll cover what is an L L M for those that are new to large language models, and we'll talk about a recent article from HubSpot, which has integrated AI into their products and what they have to say about marketers using ai. So if you're ready, let's. I posted an article on my website@gregdoig.com titled, what is A L L M? So what is an L l M or a large language model?
Well, here's a basic explanation for those new to this topic. Let's say you wanna teach a computer how to speak English. Of course, you could start by giving it a list of words in their definitions, but that won't be enough. You should also provide examples of how those words are used in sentences. A large language model is a computer program trained on massive text data. This data could include books, articles, websites, and social media posts.
The L L M learns how to use language by analyzing this data and identifying patterns. For example, an L L M might learn that the word bank can refer to a financial institution or the side of ariver. The L L M might also learn that the word cat is often used to describe a small furry animal. Once an L l M has been trained, it can generate text, translate languages and answer questions. For example, you could ask an l l M to write a poem about a cat or to translate a French article into English.
LLMs are still under development. But they have the potential to revolutionize the way we interact with computers. They could be used to create more natural and engaging user interfaces and to generate more creative and informative content. A simplified analogy might help a non-computer literate person understand how LLMs work. Imagine a person who is trying to learn a new language.
They could start by memorizing a list of words and their definitions, but this would be a long and inefficient way to learn. It would be much better if they could instead immerse themselves in the language by listening to native speakers, reading books and articles, and watching movies and TV shows. This is essentially how LLMs work. They are trained on massive amounts of text data allowing them to learn language patterns naturally.
So to sum things up, an LLM is a large language model and it's a computer program trained on massive text data to learn language by analyzing it and identifying patterns, LLMs can generate text, translate languages and answer questions, and have the potential to revolutionize the way we interact with computers.
Again, they work by immersing themselves in languages like someone trying to learn a new language listening to native speakers, reading books and articles, and watching movies and TV shows. On a recent HubSpot blog post, they talked about the use of AI and marketing, and they say it's on the rise and it is expected to reach 107.5 billion dollars by 2028.
Content creators and marketers are using AI for generating ideas or inspiration for marketing content, as well as writing copy for marketing material. AI tools like Kive, K I V e.ai and HubSpot's content Assistant AI can help generate mood boards and blog topic ideas and chat g P T can mimic a writer's style to produce press releases, social media posts, and SEO friendly blog content.
However, professionals believe AI will not replace human marketers and creators but rather assist in improving their output in producing content more efficiently. The key is for professionals to embrace AI and use it to their advantage rather than shy away from it. HubSpot did a recent survey with marketers to find out the top five ways they are using artificial intelligence. 33% said they were using it to generate ideas. 28% said they were using it to write copy.
26% say they use it to create marketing images. 25% say they use it to summarize texts into key points, and 23% say they use it to translate texts into different languages. There were a lot of stories this week about the Facebook privacy settlement. One story was posted over on Apple insider.com written by Andrew Orr on April 20th. He wrote that Facebook users can submit a claim for a share of the 725 million dollar class action settlement of a lawsuit over the Cambridge Analytica scandal.
In 2018, it was discovered that Facebook had given Cambridge Analytica, the UK political consulting firm access to considerable data from as many as 87 million users. This led to the class action lawsuit. The data was purportedly used in 2016 for voter profiling and targeting on behalf of various campaigns by Cambridge Analytica, which has.since discontinued operations.
Facebook owner meta agreed in December to pay 725 million to settle the charges of improperly using customer data without permission. It's now possible to claim a share of the settlement to receive a cash payout. Of course, that's gonna depend upon the number of people that submit claims, according to the settlement website, people in the US who used Facebook between May 24th, 2007 and December 22nd, 2022 may be eligible to receive a payment.
Facebook founder Mark Zuckerberg was required to testify before Congress over the incident and publish ads in which he apologized for the breach. So again, you can file your claim. You have to go to the settlement website, which is at Facebook, user privacy settlement.com. There's news this week that Apple is reportedly planning to release a personal journaling app code named Jurassic that will come pre-installed on iPhones running iOS 17.
The app will deeply integrate with various phone features, including location services and contacts. It will be marketed as a mental health tool, citing research that shows journaling can help with issues like depression and anxiety. The app will make recommendations to users about what to journal about based on the data stored locally on their phone, including proximity to others and daily routines. This integration could make it difficult for other journaling apps to compete.
Again, we'll have to see when the next release comes out for iOS 17. This ends this week's show. But yes, we will work to bring you more and hope you enjoyed another episode up Tech Brewed or we talk about, again, your tech brewed, just the way you like it. I'm Greg Doig, and thank you for listening.