Sponsored by
TL;DR
xAI released Grok 4.5 as a cheaper "Opus-class" rival to GPT-4 and Claude
Meta launches its own agentic image and video models, Muse Image and Muse Video
OpenAI's GPT-Live models bring real interruption and turn-taking to ChatGPT Voice
Google Photos adds Video Remix, turning ordinary clips into stylized, shareable content
Build a polished, multilingual video with Synthesia's AI assistant
Bono turns a 10-minute conversation into a blog post, LinkedIn content, and a newsletter, all in your voice
His Father Got Parkinson's. He Built Robots Instead.
Clint Brauer grew up on his family's Kansas farm. His dad sprayed the same chemicals every American farmer sprays. Years later: Parkinson's. Clint walked away from a tech career to build a different way. Today his company, Greenfield Robotics, runs a patented fleet of autonomous bots that slice weeds with centimeter precision, day or night, herbicide-free.
Greenfield is now opening shares to everyday investors under Reg A+. Reserve during Test the Waters and you lock in a 5% bonus that can grow to 20% the week the round goes live. The US has 250 million acres at stake.
Greenfield Robotics is Testing The Waters under tier 2 of Regulation A. No money or other consideration is being solicited, and if sent in response will not be accepted. No offer to buy the securities can be accepted and no part of the purchase price can be received until the offering statement filed by the company with the SEC has been qualified by the SEC. Any such offer may be withdrawn or revoked, without obligation or commitment of any kind, at any time before notice of acceptance given after the date of qualification. An indication of interest involves no obligation or commitment of any kind. “Reserving” shares is simply an indication of interest. There is no binding commitment for investors that reserve shares in this manner to ultimately invest and purchase the shares reserved of the company, or to purchase any shares of the company whatsoever.
ELON JUST PRICED OPENAI AND ANTHROPIC OUT OF THE ROOM
xAI dropped Grok 4.5 this week billing it as an 'Opus-class model' that delivers flagship performance at a fraction of the cost. It's a direct shot at Anthropic's naming and OpenAI's positioning, and the pitch isn't "we're better." It's "we're close enough, for way less."
Elon isn't trying to win the capability race anymore. He's betting enterprises are tired of six and seven figure AI bills and will take 80-90% of the performance for half the cost. It's the same playbook cloud computing used on enterprise software. Win on price, not on the leaderboard.
He's got the infrastructure to back it up too. xAI's been training on massive Nvidia clusters and pulling real-time data straight from X, which gives Grok a context advantage no other lab can copy without owning a social platform.
Whether the benchmarks hold up is still an open question. Musk has overpromised before. But the model's live now, which means devs are already stress-testing it against GPT-4 and Claude.
OUR TAKE
This is the AI market finally splitting into lanes. OpenAI owns capability. Anthropic owns safety. Now xAI is staking out cheap and good enough.
If Grok 4.5 actually delivers, this isn't a Musk problem for OpenAI and Anthropic. It's a pricing problem. And pricing problems spread fast.
Watch the next few weeks of independent benchmarks, not the announcement. That's where this actually gets decided.
QUICK HITS
Meta ships its own image and video models: Meta Superintelligence Labs just launched Muse Image and previewed Muse Video, its first in-house media generation models. Instead of prompt in, image out, Muse Image acts like an agent: it writes code, searches the web for factual grounding, and self-corrects its own drafts mid-generation. It's rolling out now across the Meta AI app, Instagram Stories, and WhatsApp, with Facebook and Muse Video coming soon.
OpenAI's new voice models actually let you interrupt them: GPT-Live (including GPT-Live-1 and mini) is powering an upgrade to ChatGPT's Advanced Voice Mode, and the headline feature is real turn-taking. It listens and speaks at once, handles interruptions without breaking stride, and free users get the mini version by default. It's a small detail that makes voice chats feel a lot less like talking to a machine.
Google Photos turns your camera roll into stylized content: Video Remix, powered by Gemini Omni, applies AI edits like cinematic relighting, watercolor, and sketch effects to your existing clips in a few taps. It's rolling out today in the Create tab for Google AI Plus, Pro, and Ultra subscribers across 14 countries, no editing skills required.
Most teams add a second database for analytics. Then manage sync, lag, and drift forever. TimescaleDB extends Postgres instead. Hypertables, 95% compression, aggregates. No pipeline. Get $1000 Credits To Start.
BUILD POLISHED, MULTILINGUAL VIDEOS WITHOUT A CAMERA CREW
One AI assistant handles the script, avatar, B-roll, motion graphics, and translation, so your team can produce video at scale.
⏱️ ~1 hour | 🔧 Synthesia account (free tier works for testing)
Why build this? Most people still picture Synthesia as "avatar reads a script." That's maybe 10% of what it does now. It's become a full video assistant that writes drafts, builds scenes, generates custom B-roll, adds motion graphics, and translates the whole thing into 160+ languages, all without leaving the editor. If your team needs training content, onboarding videos, internal comms, or explainer content and doesn't have the time or budget for real production, this is worth testing.
Steps:
Start with a prompt, not a blank page: Open a new video, pick "Prompt an idea," and describe what you need: audience, goal, key points. The assistant builds a full scene-by-scene draft, script included, in real time. You can also import a doc, a PDF, or a URL if you've already got content written.
Pick your avatar: Use a stock avatar or clone yourself from a single photo. Outfits, backgrounds, and brand colors are all customizable, and can be locked in at the workspace level so everyone's videos stay on brand.
Generate custom B-roll from the Media tab: Google's Veo 3 is built in, so you can prompt for a clip (your avatar included) without leaving the editor.
Add motion graphics from the Motion tab: Pick from 16 templates (funnels, comparisons, cycles) and it auto-syncs to your script.
Layer in interactivity if it's training content: Quizzes, CTAs, branching paths, all exportable via SCORM for an LMS.
Hit generate, then translate: One click gets you frame accurate lip synced versions in 160+ languages from the same master file.
Expected outcome: A fully produced, on brand video with avatar, custom footage, motion graphics, and interactivity, built in about an hour, with a built in path to localize it for any market without re-recording anything.
Full tutorial:
TOOL OF THE DAY
Talk for 10 minutes, get a blog post, LinkedIn content, and a newsletter, all in your voice.
Most people don't have a writing problem, they have a blank page problem. Bono skips it entirely. You talk through an idea like you're catching a friend up, and it turns that conversation into ready to publish content across every channel you actually use.
It's not just transcription either. Bono tracks your tone, your recurring themes, and who you're actually talking to, then gets sharper with every conversation. Turn on a channel and it handles publishing and performance tracking from there.
Basically a content team compressed into a conversation.
