Google just dropped a 100% free learning path on Generative AI with 9 Courses 👇
Intro to Gen AI
Intro to LLMs
Intro to Responsible AI
Intro to Image Generation
Encoder-Decoder
Attention Mechanism
Transformers and BERT
Image Captioning
Gen AI Studio
Here are NINE AI podcast tools you should bookmark and test for your next episode.
Let’s dive in 👇
Podcastmarketing.ai: This tool transcribes your podcast and then uses artificial intelligence to distill it into show notes, episode summaries, and more.
Castmagic: This AI-powered technology helps podcasters save time by transforming audio into text prepared for publication, including transcriptions, show notes, summaries, highlights, quotations, social media postings, and more.
“But large language models aren’t anything special, they’re just predicting the next token/word 😝”
That is drastically underrepresenting the power, value, and threat of LLMs.
Just because LLMs are predicting the features and probability of the next token/word/phrase does not mean it has no value or is never accurate.
GPT-4 scored 90th percentile on the bar exam…it may or may not “know” law, but that doesn’t mean it can’t get a high score, or, even outside of the exam, partially replace a lawyer’s list of duties.
There is going to be an absurd amount of B2B ChatGPT-esque apps in the next 12 months.
Here are just some of my top tips, having worked on both sides of the AI table for 10+ years 👇
Enterprise: Business Strategy 🏢
Prep your procurement teams (6+ months won’t cut it). Start small, maintain trust and transparency with end users. Consider internal-only use cases first (like Morgan Stanley and GPT-4).
Enterprise: Finance 💰
Assume it’s a paid PoC (to anchor, VC-backed startups will aim for min $100k/yr and prefer $250k/yr for production). Keep your CVC team in the loop for potential strategic investments. Budget for ML optimization. Buffer 20-40% in for trial and error.
1) Get ready to add your whole business as context for AI.
The context length war is on, sparking innovation.
These models model language and need context/examples, and Anthropic (100k) and OpenAI (32k) are both working on taking us from an article as context to full novels.
2) GPUs are a bottleneck.
Smaller models may gain traction as we await more powerful ones, also giving regulation and usage frameworks time to catch up.
Keep a keen eye on Nvidia, Intel, GCP, AWS, Azure, Cerebras, etc. - the makers and controllers of hardware and compute.
Here is a SUPER simple way to start using OpenAI APIs (with screenshots and everything!) for all of my non-technical followers that want to use more of this tech.
As a marketing manager, you can write 500 Instagram captions at once.
As an analyst, you can auto-generate memos from your insights.
As a researcher, you can quickly summarize thousands of survey responses.