AI Top-of-Mind for Feb 6
Another look at Google Bard’s new image generator. Thomas Smith writing in ‘The Generator’ evaluates the tool, which is free to use at present. As I mentioned yesterday, Bard watermarks all images, and he details just how. And as a conclusion, he doesn’t feel it is on par with either Midjourney or DALL-E3.
I’ve leveraged ChatGPT for coding assistance, but by no means is it the best for all occasions. ‘Artificial Corner’ details the best for both beginners and advanced programmers. His summary:
· When it comes to getting a grip on a new programming concept, I’d definitely use ChatGPT or Bard. They’re like info goldmines, offering a deep dive into whatever we’re curious about.
· CodiumAI is good for code testing. It’s all about speed and efficiency, giving you real-time feedback as you code, and pointing out potential pitfalls.
· GitHub Copilot is my go-to general-purpose coding assistant.
· Experienced programmers can also benefit from coding assistants like AWS Code Whisperer and Tabnine.
Have you had a bad day with Google search? You are not alone! Ronke Babajide in ‘Bouncin’ and Behavin’ Blogs’ makes the following points:
· AI-powered content mills are generating content at a speed that no human writer can match.
· And that spammy, low-quality content is optimized for Google SEO by AI in a way that outranks any human-written content.
· If you click on the top search results these days, you’ll find low-quality content that is often riddled with errors and false information. But the Google algorithm can’t recognize that.
· As a result, content written by robots for robots is clogging up our search results.
On the educational front, Cassie Kozyrkov pulls apart the various phrases and acronyms we run across, even in my postings. If you are curious about embeddings, vector databases and search, and even k-NN and ANN, read-on. Being new to this, I’m just running across some of her earlier postings, which are quite instructive. The article includes a few links.
Next, Mohit in ‘Medium’ provides his view on the five ‘coolest’ AI projects. They span QR code generators, a tool that coverts PDFs to chat, AI-generated emojis, an X bio generator, and finally a headshot enhancement program.
Turning to retail, Amazon is the latest to bring AI-assisted shopping to their site. ‘Mobile Marketing’ details‘Rufus’ which sounds like a dog. From the article:
“With Rufus, customers are now able to shop alongside a generative AI-powered expert that knows Amazon’s selection inside and out, and can bring it all together with information from across the web to help them make more informed purchase decisions.”
And finally, over to the workplace, ‘Ad Age’ looks at which jobs we should outsource to AI, leaving time for true creativity. The nine tasks:
· Photo/video retouching
· SEO/paid search headlines
· Pharma ads
· RFI/RFPs
· Media plans
· Contracts
· Trafficking
· Digital accessibility
· Research
The message is: Scale output, not input. Not only do we need to give people the power to produce more, but we also need to free them up to devote energy and time on things that require focused expertise. If you save your human insight, skill and time for the parts you need to create, you’ll make more compelling things.
