Texti Newsletter #5

Christmas is around the corner, this means everybody's wearing a funny Christmas sweater, and everybody's rushing to complete all the project in time. Even if you need to cut the corners.

The highlights of the week:

  1. Pika 1.0 is already showing off
  2. Google lied in your face!
  3. Hint of the week.

Pika 1.0 is officially the new head turner!

If you didn't know what Pika is, it's the new kid on the block! It turns head like no-one else. You tell Pika: time lapse photography of a rose bud unfurling, blooming, red petals expanding

You tell Pika: a lion fish swimming underwater

Pika: a cinematic film still of a breaching whale, falling down into the water, splashes, sea foam

Well isn't this impressive? It sure is, but that was to be expected. A video is just a series of pictures following a little progression with each frame. It was just a matter of time until we learn how to make pictures progress. To be completely frank with you I think it's not the how to do it, but rather drop the cost.

Currently image generation still costs a fortune if you're not a billion dollar company. NVIDIA is selling their hardware for hefty price tag and it's still heavily unaffordable.

Yet everybody's predicting that the next year it will be the year of Text to Video, which makes everyone super excited, but at same time the price tag for this will be pretty astronomical.

Gemini Demo was all a lie!

Last week, Google Announced Gemini, we've talked about it in the last Newsletter. We did say that the Demo was mighty impressive, and I still stand to this words! It was really impressive, so impressive in fact that it raised questions to many!

To be fair, Google said at the end of the document where they presented the demo, that there's an in depth analysis of the demo -- in a separate document how they actually have build every piece of the demo, however, the demo looked flawless and all in real time.

Now everybody got really disappointed that this presentation box, was kept together with tape and not actual structure.

But worry not, the internet is broad, there are a bunch of nerds out there who can react promptly to this.

Here's one: A guy called Greg, just remade the Google Demo in real time, with OpenAI. This is a closer representation of the current state of AI.

It first needs time to analyze and process the information it has, process the images, convert it into text, and then convert it into speech, hence, you can see in the video that there's still a substantial delay in the communication between the AI and Human.

Still is very impressive, and the days were life's is going to be visually assisted is really close.

You can watch it right here:

video preview

For now we can still say OpenAI is the one and only King of the current race! Sam Altman is still holding that crown really tight on his head!

Hint of the week 💡

If you are struggling with the AI responses, and AI still can't give you a proper answer, there's a simple solution. Split the task into smaller pieces, and feed them one by one into the chat bot.

AI is dumb, even if it seems smart, complex tasks are hard to manage for him, as it's lacking background knowledge, and can't see the big picture. If you feed it granular tasks, it will accomplish them with large success.

That's it folks, see you next week ❤️️️️️️️

Happy Prompting!

Remember to invite your friends to subscribe at https://texti.app/newsletter

{{ address }}
Unsubscribe · Preferences

© 2024 Texti Newsletter