sarahai is an autonomous ai assistant that works through whatsapp. $9.99/user/month .
the founder has 20+ years of tech experience — early mobile payments acquired by google, deep learning at ibm watson .
the technical story: they've built orchestration layers on top of llms to handle non-deterministic systems, voice pipeline latency on whatsapp, and token cost management .
users say it's replaced their virtual assistant. one ceo running two startups said she just drops voice notes and it schedules everything .
the adoption pattern: week 1-2 for daily briefs and voice-note task capture. week 3-4 delegating team tasks. month 2-3 it becomes the operational layer .
for creative professionals: an ai that runs the operational side of your business means more time for the creative work
agentic video understanding is one of those features that sounds simple until you realize what it can actually do
being able to reason through an entire video instead of just grabbing frames could make a lot of workflows way less painful
the explainer is worth watching 👀
yesterday, we released agentic video understanding with Gemini.
here's a great video explainer from @MaartenGr on how it works:
https://t.co/7efC3CWPQl
@openclaw@github openclaw 2.0 is here. now the interesting question is how the hell they got it this far 👀
scaling an open source project that fast while keeping security in check is the part worth digging ito
@github@openclaw@steipete the intersting part is everything that had to happen behind the scenes to get an open source project scaling that fast without turning security into an afterthought
that's the part i'd want to hear about 👀
@FilmUpdates danny boyle + jack o’connell + claire foy?
yeah okay, i'm watching this
thjanuary release makes it even more interesting. curious what kind of movie this turns out to be.
$7 vs $90 vs $150 per seat
the jump from starter to professional is doing a lot of heavy lifting here 😭
i'd definitely want to test the actual workflow before paying that much.
ai video is getting way too close to an actual editing workflow
google’s new gemini omni 1.1 flash adds 4k, scene extensions and first/last-frame control
as someone who watches way too many movies, the amount of control here is getting interesting fast.
claude code users are getting a permanent 25% bump to weekly limits starting september 14
the temporary 50% increase stays until then
honestly, i’ll take the extra usage.
Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place.
@kimmonismus anthropic deleting the tweet and then clarifying the limits is definitely one way to handle it 😭
at least the permanent 25% increase is still a pretty solid upgrade.
@NotionHQ honestly, the tiny updates are sometimes the ones you notice most
a bunch of little 2% improvements can completely change how a tool feels day to day.
@Variety $3m on friday for the dog stars
rough start considering the cast and ridley scott behind it
the premise sounds like exactly the kind of movie that could find its audience over time though.
@FilmUpdates anne hathaway joining tom cruise for days of thunder 2
okay, now i’m paying attention. this cast alone has me curious what they’re cooking.
big move honestly, cutting off model access two months out feels less like a clean breakup and more like openai making sure a competitor's tool can't keep running on their infrastructure. rough timing for anyone who built their whole workflow around cursor + gpt specifically 👀
grok bot can shop for you now
connect link, tell it what to buy, and let it handle the checkout
we're getting dangerously close to outsourcing the entire “add to cart” experience 😭
@GoogleDeepMind@FlowbyGoogle controllable video generation is where these tools start getting really useful
faster iteration + more control means less time fighting the model and more time actually making the thing you had in your head.
@yacineMTB the idea of ai models becoming the layer that abstracts away different hardware is pretty interesting. if developers can stop thinking so much about the underlying accelerator, distribution becomes a huge advantage.