Langchain secures $10M seed round

The popular development library which makes it easier to create AI powered experiences by connecting together different AI services, has raised a $10M USD seed round. Langchain was released just a mere 6 months ago! A testament to just how fast the industry is moving and new solutions are being developed. [read more]

Google I/O coming up, what to expect

Google I/O kicks off at 10am PT tomorrow, returning to in-person for the first time since the pandemic. It’s safe to say that Google was caught off guard and has been playing catch-up since OpenAI released ChatGPT 6 months ago (Nov 30, 2022). With their release of Bard being to somewhat mixed reviews and with a bit of controversy, we expect Google will likely announce announce an update to Bard as well as numerous AI integrations with their productivity apps, such as Google Docs. Whatever happens, I’m sure that tomorrow will be full of exciting news and a lot of details to sift through. [watch online]

IBM launches new AI services

Hot off the heals of announcing that they will be using AI to replace 7800 jobs, IBM is back in the news with WatsonX, a new AI platform aimed at the enterprise market. For the most part, it appears that IBM is just releasing the same AI models & services already available from Microsoft, Google, and Amazon, but IBM claims that what makes their models unique is that they have been trained specifically using enterprise data. If that’s a compelling story is yet to be seen, but for now you can add IBM to your list of AI provider choices. [read more]

What exactly is ChatGPT?

Put simply, ChatGPT is a chat program (a webpage) which interacts with a Large Language Model (LLM), specifically the GPT-4 LLM. You may have also heard of the term NLP or Natural Language Processing as well. That is the ‘chat’ part of ChatGPT which allows you to communicate with it…naturally. When you type a question (aka prompt), that text is sent to the LLM for processing and the response is output in your chat.

So you might now be wondering what an LLM is. I’m glad you asked. An LLM is a computer program that has been given (trained) on A LOT of textual information… documents, web pages, etc. It can recognize patterns in the text, complex patterns. Just like a child that has been taught their ABCs, if you start to sing A..B..C.. they will start to sing the rest of the alphabet. Likewise, given some text, a prompt, the LLM will ‘predict’ what it thinks comes next. Of course the patterns known to the LLM are incredibly complex, so it appears that it’s responses are intelligent, when in reality it has no understanding of what it’s responding with. It is merely guessing what it ‘thinks’ should come next. Just like the predictive text on your phone when you’re typing a text. That’s it. :)

In future editions we’ll cover a little more about GPT and what makes it tick. Stay tuned.