What’s up team!

Today we’re learning why all of a sudden there are 10232 products applying ChatGPT, why prompt engineering matters and how to supercharge your video editing skills.

Let’s go!

🗞 News in 30 secs’

🇬🇧 The UK have announced a £100m investment on a UK GPT
Being Dubbed `Sovereign AI` by Rishi Sunak, the model is tasked with making the UK "globally competitive," and will work with the industry to make these systems safer and more reliable. Coming from the same government who lost under-reported on covid vaccinations due to an excel error… we’re skeptical of seeing anything major soon.

😅 Generative AI is transforming Hiring.
The BBC reported on how AI is transforming hiring. The article focusses on written materials composed via generative AI, but somewhat misses the real opportunity. AI is already being deployed to help train interviewees and remove bias from the hiring process in real time. That’s a bit more impactful than an AI cover letter.

🎶 The Generative AI music debate continues 

And this time it’s Grimes letting us know that if you want to use her voice in your tunes she’ll share 50% of the royalties with you. The announcement comes as universal crack down on the Drake track released without permission last week.

⌚️ AI Wearable Tech

Former Apple employees unvailed Humane. It’s an AI assistant that you wear, that can do things. Probably shows the direction of travel but nothing mind blowing just yet - we’ll keep 👂️ and 👀 .

Prompt Engineering 101

👉️ Prompt Engineering is basically the difference between you asking your mum for pocket money when you’re 7 vs when you’re 10. When you’re 7 your inexperienced and ask straight up for the cash. By 10 you’re smart. You smchooze with some chores first, maybe make your bed then ask sweetly for a specific amount.

👉 Prompt Engineering is similar, asking the same question in different ways but getting drastically different outputs.

👉️ Prompt Engineering matters now because we’re all learning AI. If you know how to get the best information back then that alone is valuable for businesses or the internet.

👉️ Prompt Engineering matters tomorrow because the current #1 restriction on LLM’s are the sheer cost of each question. At scale that get's wild. Prompt engineering gets you too the desired response in less questions.

👉️ Effective prompts == better product and cooler finds.

👉️ Yes. There’s already a job board for that. Prompt-engineering-jobs.com has a bunch of jobs posted already including Remote Cannabis KI-Prompt Engineer.

👇️ Harvard doesn’t have a AI Prompting course yet. In the meantime, check out some sweet sweet prompts here 👇️ 

What’s Appnin’

veed.io is like Canva for video editing.

From zero to hero in a minute or two.

Four cool features we used:

• Automatic subtitle generation which are automatically synced. You no longer have to spend hours transcribing videos.

• Background noise remover that cleans up unwanted background noise from videos.

• Automatic green screen remover - rendering this use to take hours in After Effects, now it takes minutes.

• Text-to-speech option that turns text into speech using AI-generated voices.

Also add funky fresh sound waves 🌊

Adding these same effects in Adobe takes alllooooot of training and alooooot of cash. Now you can do it on Veed in a few minutes.

It makes us hopeful for the accessibility space too, soon subtitles in every language will be a default. Heck. Why not just morph every actors mouth to have them realistically speak every language?

So Why AI Now?

If you’re like me, you’re thinking..why is AI Popping off now.

What made 2023 the moment when we jumped from barely abstract pictures generated from text to the pope looking fresh in some.

(Old news now... 👇️ )

Well it’s basically down to two factors.

1/ The amount of data we have at our disposal has grown like wild fire in the past few decades. What used to be written and logged on paper or print outs is now stored in the cloud.

2/ Better and better models. Seems obvious right. Well even if you had all the data in the world, researchers found that traditional AI models plateaued at a certain point. Then Deep Learning and Neural Models came along and said..hold my beer 🍺

More data is always better. Even with the same amount of data, larger neural networks give better results and take longer to plateau.

Hardware has also kept up, with companies creating more and more complex hardware that can actually perform the calculations needed to create these networks.

And what precisely is a neural network? Let’s get our beer back and learn that tomorrow.