ICYMI: Grok 2.0 mini supports LaTeX formulas 🤖

A new image gen model drop next week 👀

This is Imagen3 prompted to show one of my posts on the monitor. Next week, you will be generating such images with perfectly written text from the first attempt.

ICYMI: The SearchGPT waitlist is now closed. In case you have applied already you should be good 👀👀👀

All SearchGPT prototype spots are filled. Thanks for watching 👀👀👀

🚨 BREAKING: Perplexity now supports 2 new image gen models: Flux.1 by BFL and Playground v3

Both options are now available in settings and can be used for image generation for Perplexity Pages.

In the image generation UI you can also specify a certain image style separately 👀

And one Flux example 🔥

Playground now has its new v3 model available to all users on the main website in beta.
https://playground.com

ICYMI: Gemini update rolled out with smarter AI for complex tasks
https://www.testingcatalog.com/icymi-gemini-update-rolled-out-with-smarter-ai-for-complex-tasks/

🗞 #gemini

ICYMI: OpenAI closed SearchGPT waitlist after recent invite waves
https://www.testingcatalog.com/openai-closed-searchgpt-waitlist-no-more-early-access-slots-available/

🗞 #chatgpt

Perplexity adds Flux.1 model for Pro users alongside Playground v3 update
https://www.testingcatalog.com/perplexity-adds-flux-1-model-for-pro-users-alongside-playground-v3-update/

🗞 #perplexity

🚨 BREAKING: Runway is working on an option to "Extend" videos, and it seems that it will be possible to generate 20-second movies.

or is it 10+20 seconds? 👀

The feature is still in development and might not yet be ready. But there are more 👇

Previously discovered Motion Brush and Camera Control seem much more polished now. Besides that, Runway is working on supporting different aspect ratios including Portrait.

This might be a huge deal as you will be able to easily post these generations on TikTok and Reels 👀

As many of you probably have heard, LumaLabsAI is experimenting with its new v1.5 model which is available to selected creators.

For model evaluation purposes, creators can access a new tab called Rate to help evaluate output results.

Lmsys arena but for videogen 👀

Runway’s potential upgrades could bring 30-second AI videos
https://www.testingcatalog.com/runways-potential-upgrades-could-bring-30-second-ai-videos/

🗞 #runway

AI week begins: 1st in line - Dream Machine 1.5 AI video gen model by Luma Labs 🧪

Queuing! Saved all my 30 free generations for this month for this exact moment.
Try here https://lumalabs.ai/dream-machine/creations

Testing Mystic Alpha today and this picture says about my experience better than any leaked UI screenshot.

j/k UI screenshot below 👇

The mystic image gen model is available on Freepik as Alpha. Currently requires an invite. In general, it is quite comparable to Flux and it is also able to generate text very well 👌

I was fast ⚡️ Got my Luma 1.5 generation before it got overloaded. Great result from the first attempt!

Image gen - Mystic Alpha
Video gen - Dream Machine v1.5

"Cyberpunk hacker robot".

UPD: Nope - this is still V1, need to wait until morning. V1.5 only works with text prompts.

Perplexity now has better programming tooling suitable for a wider range of tasks 👀

But in case you ask it about stock prices, it will do double work along with the Tako data fetcher. Slowly rolling out!

Google just doesn't stop.

28 of August has been added to the list of upcoming announcements, committing +1 announcement to the already existing list of 5 planned earlier.

The end of August is planned to be hot 🥵