Here I wrote down some notes after watcihng the Intercom Off-script series about how they are facing the AI age and how the see it playlit
Video #1
It will blow up the total address of all Market because it will and I will customers or users who know what they want to do however they don’t know how to do it for instance in Excel or monday.com, salesforce who are great product very robust products but not so many people know how to use them, AI will enable them to use them at its maximum
We need to rebuild our products our societies around this new technology and it is day one
Video #2
Being disturbing pollution change and reshape our society because before that everything was made by hand and after that breakthrough moment we were able to have houses cars clothing
Also technology is like a one-way street that once you go and cross that door you never look back before for instance once we had the iPhone in our hands we did not think about the traditional phone call or mail post
What matters the most is that customers and businesses get greater experiences, it boils down to it
Video #3
Attention algorithm and then wrap up with a Transformer machine learning technique Then the Transformer was trained with the huge amount of data that never before And finally it seems like it is understanding and it is capable to do some summarizing words follow some instructions
An example of what AI needs was when governments increase their budget in order to roll out new highways because they saw that as much highways and streets interconnected you have the more trade Commerce and tourist activity you will get in this case AI needs more compute infrastructure and so on (Things happen fast when the value is there and apparently AI is delivering some value)
Apparently it is learning because it is paying attention to the important things
It’s got some limitations but you can’t work around those limitations
People say that it is just a talking prediction like predicting the next word on the sentence
RAG Is given the llm more context in order to make it understand the current situation therefore it will provide a more accurate answer
Take a small series of tasks That you can really over-delivery on, not too broad.
We have used openAI models as building blocks, Maybe five or 10 different prompts Underneath the hood, Each of them doing a specific task like search for this customer in the db, then get their last order, then disambiguate this question
From an engineering perspective take five things your product should do and isolate them And get AI to do that task separately
Video #4
Mother sauce and software in general seems to be very complicated and clunky because they serve a broad amount of users, It’s like cities in general they are designed for a general audience not for cyclist or pet owners or car drivers
Now with AI you can have UI design that is generated on the Fly based on the user request instead of having a form where users submits email phone number number order it will go and retrieve those data from the database and make up a new form just to confirm that the information is correct and it will be completely Custom Tailor to each requested and even to each user
It raises the floor for the designers, the human touch becomes more valued
If you wear a drummer in the 80s and you see that a rhythm controller has come out you would thought that you are cooked but it happened all the opposite more drummers came to play their songs (democratizacion of music) So you don’t need to become a master at your craft in order to make music
however there still need of people who are a master at the craft even when AI provides shortcuts
Embrace the squiggle (process of designing something, converge, diverge ideas, discussions, etc)
Video #5
Better, faster cheaper on online business
Customer service in businesses prior the war was more personal they made individual transactions, and then as the business expanded they needed to hire more customer service in order to make it more Global more professional and then after that the 01800 phone lines came up and that was the time that the call centers started then with computers and internet the email and then the CRM too, chatbots, etc, also how customers leave feedback from the service they got and that was a time where you needed a quality customer service.
Easiest decision is to cut people in order to make profit…
So in order to increase the quality of your customer service at the internet scale is very difficult
Video #6
Big boys and small are no longer selling software seats to humans, but software doing the work.
We don’t pay for CS, we pay what’s in the receipt, and CS is all around.
Having a internet business bring the benefits of scales which is selling around the clock, around the globe
Heretofore (before now, new word)
Agents gen1 (conversation, text boxes) and gen 2 (more familiar, proactive and take action)
Cannibalize themselves if they want to survive and Netflix did with their shipping DVD business before streaming.