In partnership with |
 |
|
|
[Webinar] Can you prove AI is working? |
|
AI is in your engineering workflow. While the token spend shows it, the throughput doesn't. The human is very much still in the loop, and that's a context problem. |
Join live on Aug 19 (FREE) to see: |
The 4 metrics to measure the gap where gains leak out before production. The 8 stages of context maturity, the specific walls capping your metrics, and a free tool to pinpoint where your team is Why more MCPs and bigger context windows aren't enough and what it takes to get real value from your agents. |
Register Now |
A new AI video model just did something no closed competitor can match: it renders a 10-second clip in under 7 seconds - faster than the clip plays back. LTX-2.5, built by Lightricks spinout LTX, is also free and open-weight, meaning anyone with a decent GPU can run it today. |
That combination - real-time speed and zero cost - is a real threat to every subscription video tool charging by the render. If open models can now match or beat the paid leaders on speed, what's left for closed platforms to sell? |
Today in AI: |
LTX-2.5 renders AI video faster than real-time xAI's Grok Bot works your apps around the clock Claude quietly watermarks everything it creates |
|
|
|
What's new? Lightricks spinout LTX released LTX-2.5, an open-weights AI video model that turns a still image into a 10-second, 720p clip in 6.8 seconds - faster than the clip takes to watch. The 22-billion-parameter model runs on a single high-end GPU and ships free for anyone to try today. |
What matters? |
It's actually free to run. The weights are live on Hugging Face and inside ComfyUI, with no per-clip fees or watermarks for organizations under $10 million in annual revenue. Speed comes from smarter compute, not less of it. A new decoding technique called Diffusion Fidelity Rendering pours extra processing into complex scenes - crowds, fast motion - instead of spreading it evenly across every frame. It's already the most-downloaded open video model. LTX says its model family has passed 33 million downloads, ahead of every other open "world model" on the market. |
Why it matters? |
Every closed video model still makes creators wait, queue, or pay per render. LTX-2.5 collapses that wait to less time than the clip itself takes to watch, and hands the weights to anyone with a decent GPU - the kind of gap that pulls hobbyist video creators away from subscription tools fast. |
|
GUIDE |
|
|
|
xAI launched Grok Bot in public beta, an AI agent that gets its own cloud computer, signs into the apps and tools you already use, and works through tasks around the clock - checking in with you only when it needs approval. |
What matters? |
It learns by watching you. Grok Bot picks up a workflow the first time you demonstrate it, then repeats it going forward without being told again. You run it like a group chat. Multiple Bots can coordinate with each other in parallel, splitting up work the way a team of coworkers would. It's live now, in beta. Sign up at x.ai/bot - access starts with SuperGrok Heavy, Cursor Ultra, and Cursor Teams Premium subscribers, with an enterprise waitlist open. |
Why it matters? |
Every AI teammate promise so far has needed constant supervision to be useful. Grok Bot's bet is that watching once is enough - if that holds outside the demo, it pressures every other agent platform still asking users to write instructions by hand. |
|
SPONSORED BY ROKU |
How Jennifer Aniston's LolaVie brand grew sales 40% with CTV ads |
|
The DTC beauty category is crowded. To break through, Jennifer Aniston's brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category. |
Learn more |
|
|
|
Claude now embeds an invisible watermark in everything it generates - text, images, code, and files - a move Anthropic confirmed as part of its commitment to the EU AI Act's transparency rules, and one TechCrunch reports is already dividing developers. |
What matters? |
The mark travels with the content. Anthropic says the watermark survives copy-and-paste and persists through some editing, so text pulled from a Claude chat into another doc still carries the signal. Images get a different tag. For file types like .svg, .png, and .jpg, Claude attaches signed provenance metadata instead, which can also flag if a file was tampered with after export. Not everyone's convinced. Forbes reports the move has sparked backlash over privacy and centralized control, even as supporters call it responsible transparency. |
Why it matters? |
For an audience making AI art and AI-assisted content, invisible provenance tags are about to become part of the toolchain whether creators want them or not. The upside is fewer arguments about what's "real" - the downside is a permanent, hard-to-strip record of exactly how a piece was made. |
|
Everything else in AI |
River AI raised $1.1 billion across seed and Series A rounds, backed by Nvidia, AMD, and Y Combinator, to help developers customize open-source AI models in minutes instead of weeks. |
Nvidia partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to mobilize over $500 billion in AI compute infrastructure financing. |
Unsloth lets you run and fine-tune open image, video, and audio models locally through a free desktop app - no coding, and no cloud bill. |
Want more out of your AI stack? Grab the Complete AI Bundle - prompts, workflows, and guides built for people who actually ship with AI. |
|
Essential AI Guides - Reading List: |
|
|
Let us know! |
What did you think of today's email?
Before you go, please give your feedback to help us improve the content for you!
|
|
|
Work with us |
Reach 100k+ engaged Tech Professionals, Engineers, Managers and decision makers. Join brands like MorningBrew, HubSpot, Prezi, Nike, Ahref, Roku, 1440, Superhuman, and others in showcasing your product to our audience. Get in touch now → |
|
