Eliana Skye ๐๐ retweeted
bernard arnault fired @fabricegrinda from his own company at 26.
he started over on $100K of credit card debt, missed payroll 27 times, lived on four cups of ramen a day, and sold his next company for $80M at 29.
then he built OLX to 300 million people a month and became one of the most active angel investors in the world: @Uber, @Airbnb, @stripe, @AlibabaGroup, 1,300+ startups.
he also sold @TencentGlobal at a $400M market cap and passed on @OpenAI. "i left more money on the table than i've made ever since."
i asked him about offering jack ma $10,000 for the alibaba domain, the friday 5:01 PM trick @eBay never caught, why AI won't take your job, and what asking out 1,000 strangers taught him about rejection.
@fabricegrinda on nontechnical. dinner in new york, 79 minutes.
00:00 fired at 26, an $80M exit, 1,300 startups
00:57 nice: a PC at age 10
04:38 princeton: no friends, all A-pluses
08:12 why he went to mckinsey (and why you shouldn't)
11:53 offering jack ma $10,000 for the alibaba domain
15:36 "when i stopped being judgmental"
16:07 ebay offers $20M. he has 100 euros.
17:35 "what's the worst that's going to happen?"
18:15 how bernard arnault ended up with 50.1%
21:28 fired from his own company
22:59 a 20 million euro tax bill on zero income
25:45 zingy: he had never listened to a song
28:32 missing payroll 27 times
30:54 four cups of ramen a day
31:44 a $40M offer becomes $80M in cash
33:37 OLX: launching in 100 countries at once
37:53 the friday 5:01 PM trick ebay never caught
41:16 pitching the castros a free trade zone in cuba
42:48 up to $10 billion for craigslist
44:45 FJ labs: the four things he looks for
47:17 the uber memo: "i'm going to regret this for the rest of my life"
50:08 selling tencent at a $400M market cap
52:11 passing on openai
53:30 "whatever you do next, you get a check"
54:35 his two AI agents
56:06 is software engineering dead?
57:13 a $20,000-a-month AI bill, and the land it found
59:05 life is a video game
1:01:57 the parenting advice
1:03:30 "one of the happiest persons that's ever lived"
1:08:00 the meaning of life
1:09:11 will AI take all the jobs?
1:13:07 asking out 1,000 strangers
1:18:48 "things are pretty epic"
Sometimes a scene looks perfect on its own but doesn't make sense when placed next to the previous one.
Flova Agent can review the project as a whole, identify these disconnects, and help complete missing moments so everything flows more naturally.
Context is just as important as generation quality. flova.ai/?refCode=2KEYEZ6G
Eliana Skye ๐๐ retweeted
Backed by YC and tested by NBA star Moses Moody, our glasses tell you when your brain peaks.
Your ring tracks your heart. Your watch tracks your sleep. Nothing tracks the eyes and brain running the show.
STASIS reads your 100,000+ daily eye movements to tell you:
โข When youโre at peak focus
โข When your eyes start burning out
โข When scrolling and blue light starts costing you sleep
6 sensors. 11mm frame. 92% accuracy. No external cameras.
Pre-orders are open. Get yours.
wearstasis.com/checkout
Eliana Skye ๐๐ retweeted
More ways to create, plus a major credit discount.
Dreaminaโs Black Friday lineup includes REAL unlimited generation with top models and up to 83% off credit purchases. The unlimited access periods vary by model.
Dreamina Black Friday Sale is Live!
๐ Unlock UNLIMITED access to our video and image models.
๐ฌ Unlimited Video Generation
> Flagship Seedance 2.5: up to 30 days
> Seedance 2.0 series: up to 30 days
> Unlimited generation of videos up to 30 sec at 1080p
๐ผ๏ธ Unlimited Image Generation
> Seedream 5.5: up to 60 days
> GPT Image and Nano Banana series: up to 60 days
> Seedream 5.0 Flash: up to 365 days
๐ฅ Bonus:
๐ฃ Credit Purchase Sale: up to 83% off - best deal of the year.
๐ Limited-time offer. Stock up for a full year.
๐ Special Giveaway: 30 lucky winners will each win a one-year Advanced Plan!
How to Enter:
1. Repost and follow @dreamina_ai.
2. Comment below and share your experience using our features (motion studio, marketing studio, drama studio... )
Don't miss out. Explore our latest features and create more with Dreamina!
#DreaminaAI #DreaminaSeedance2_5 #blackfriday #DreaminaSeedream5_5
Eliana Skye ๐๐ retweeted
The Dreamina Black Friday offer combines more generation with a credit deal.
Explore REAL unlimited access to top models for set periods, and save up to 83% on credit purchases. Check the offer page for the available models and terms.
Dreamina Black Friday Sale is Live!
๐ Unlock UNLIMITED access to our video and image models.
๐ฌ Unlimited Video Generation
> Flagship Seedance 2.5: up to 30 days
> Seedance 2.0 series: up to 30 days
> Unlimited generation of videos up to 30 sec at 1080p
๐ผ๏ธ Unlimited Image Generation
> Seedream 5.5: up to 60 days
> GPT Image and Nano Banana series: up to 60 days
> Seedream 5.0 Flash: up to 365 days
๐ฅ Bonus:
๐ฃ Credit Purchase Sale: up to 83% off - best deal of the year.
๐ Limited-time offer. Stock up for a full year.
๐ Special Giveaway: 30 lucky winners will each win a one-year Advanced Plan!
How to Enter:
1. Repost and follow @dreamina_ai.
2. Comment below and share your experience using our features (motion studio, marketing studio, drama studio... )
Don't miss out. Explore our latest features and create more with Dreamina!
#DreaminaAI #DreaminaSeedance2_5 #blackfriday #DreaminaSeedream5_5
Eliana Skye ๐๐ retweeted
Within a week of launching the Voice Arena Dictation Bench, we have a new #1 ๐
Since launch, 20+ companies from around the world have asked us to evaluate their models on the benchmark. Today, weโre adding 5 of them.
@aquavoice debuts as the new #1 with a zero edit rate of 80.5%, with strong showings from Blurt by AssemblyAI, Willowโs new model, Spokenly and Kivi.
Moving further, weโre also open-sourcing part of the benchmark on Hugging Face, so anyone can see what the data looks like and reproduce the scores themselves.
More models will be joining the board next week.
Eliana Skye ๐๐ retweeted
Frontier AI has solved math. Science. Coding. But it still hasnโt solved taste. Until now.
Introducing Arcade 3.0: the AI video playground for marketing teams.
Type one line, answer two questions, and Arcade does the rest.
Voice, captions, and music included.
Arcade runs its own harness on top of the best models.
Itโs seen your actual product AND knows your brand.
The numbers show it: in a blind test, B2B marketers chose Arcadeโs video 2x as often as videos from Claude Opus 5.5 and Gemini Argon.
And you can change one thing without redoing the whole video.
Oh, and this video? Made with Arcade.
Comment "ARCADE" in the next 48 hours for 50% off your first 2 months of Arcade Pro.
Eliana Skye ๐๐ retweeted
Notetaker is live today on iPhone. Now, your meeting notes can go wherever work takes you, for free.
Use Notetaker on desktop for virtual meetings and on iPhone when youโre meeting in person.
Your notes stay together across devices, so you can pick up right where you left off.
A lot of AI products focus on the intelligence layer.
But the execution layer deserves just as much attention: where the agent works, what it can access, and how it delivers the result.
๐๐ง๐ญ๐ซ๐จ๐๐ฎ๐๐ข๐ง๐ ๐๐ฒ๐๐ฅ๐๐ฐ ๐.๐
Raised AI teammates with tools built in โ not a blank agent you babysit.
โข๐๐ ๐๐ง๐ญ ๐ฐ๐ข๐ญ๐ก ๐๐ง๐ฒ ๐ก๐๐ซ๐ง๐๐ฌ๐ฌ โ Claude Code ยท Codex ยท Hermes ยท Pi ยท OpenClaw โ shared context.
โข ๐,๐๐๐+ ๐ฅ๐ข๐ฏ๐ ๐๐๐ญ๐ ๐ฌ๐จ๐ฎ๐ซ๐๐๐ฌโ Not a scavenger hunt for APIs.
โข ๐๐ฅ๐จ๐ฎ๐ ๐ฌ๐ฎ๐ฉ๐๐ซ ๐๐จ๐ฆ๐ฉ๐ฎ๐ญ๐๐ซ โ up to 64 vCPU ยท 128GB RAM ยท 1.28T storage
โข ๐๐ง๐ฒ ๐๐ก๐๐ง๐ง๐๐ฅโ Browser, Slack, Teams, Telegram, WhatsApp
Ready in ~30 seconds โ myclaw.ai/new-launch
Eliana Skye ๐๐ retweeted
Weโve raised a $7.5M seed round led by Felicis so AI can read documents as easily as it reads databases.
Finance, law and healthcare run on documents, and not databases like any other industry.
Soon AI agents will read and write more documents than humans, and theyโll need to query them the way software queries tables.
Weโre building the computable layer that makes this possible.
To do this at scale, we've brought together ML researchers and engineers from Google DeepMind, Cognition, and quant trading firms.
We use computer vision and dual-stream models to turn complex multimodal documents into formats LLMs and AI agents can understand and reason over.
If your team wants to turn messy documents into reliable, AI-ready data, book a demo:
unsiloed.ai/book-demo
can you tell which one is real?
84% couldn't
introducing Avatar 6, #1 on New Measure's Avatar Arena
what's new:
realism - passes for the real you
context-aware gestures - matches what your script means
reference - up to 30 secs of video + a photo
try it now โ
If you're still editing every Black Friday video from scratch, you're making life harder than it needs to be.
A more organised timeline can make revisions much easier. Multi-Track Editing lets you keep your main footage, B-roll, captions, music and sound effects on separate tracks. You can adjust individual elements without disturbing the whole project.
For anyone looking to try it, CapCut Desktop Only is offering 90% OFF your first month for eligible new users, users nearing subscription expiry, and lapsed users whose subscriptions expired over 30 days ago. The monthly subscription starts at $3.99 or less. Select eligible new users may qualify for a free trial. Prices vary by region.
You can also use CapCut Agent EditPilot to make timeline adjustments through simple conversation.
And for additional visuals, Seedance 2.5 4K can generate clips up to 4K through AI Video or Video Studio in select regions outside the US. US rollout coming soon.
Give this setup a try:
tinyurl.com/2v23rn22
#CapCutPro90Off #CapCutPC #CapCutDidThat #ChatwithCapCutAgent #CapCutMarketing #CapCutSeedance254k
1,730 tasks in 24 hours. Six robots. And not all of those tasks were the same. ๐คฏ
During AGIBOT's Shanghai convenience store trial, the G2 robots handled 683 customer-service interactions, 369 product pickups and deliveries, 147 coffee orders, 126 breakfast servings, 277 shelf inspections and restocking tasks, and 128 table-cleaning tasks.
Think about what that involves. Recognizing a product on a shelf, picking it up without knocking over nearby items, responding to a customer, then switching to an entirely different task. Each one puts different demands on perception, decision-making, and physical control.
The robots were also tested across lunch, evening peak hours, overnight operations, and the morning rush.
That's what makes the 1,730 figure worth looking at: it's not one movement repeated 1,730 times. It's a mix of real retail tasks in a changing environment.
#AGIBOT #EmbodiedAI #Robotics
A new generation of faster video creation is here! ๐
Try Seedance 2.5 Flash on Flova and enjoy 40% off. Spend less time waiting for generations and more time exploring your creative vision.
Your workflow just got faster. flova.ai/?refCode=MNZFK6P6
you shouldnโt need to build a billion dollar company to get a world class investment bank to take you seriously.
introducing @Preambient. the autonomous investment bank for founders.
we help you raise capital, acquire your competitor or sell your business.
We get the deal done 4x faster at half the fees, fully outcome based.
our team was trained at Goldman Sachs, JPMorgan, Morgan Stanley
our ambition is to become the infrastructure for private market transactions. the platform where private businesses are priced, financed and traded.
Can we talk about the cost of iteration? Flash is $0.152/sec for 720p text-to-video and image-to-video. Around 34% less than Seedance 2.5 could matter across a whole campaign.
More ideas to try. Less time waiting. โก
Introducing Dreamina Seedance 2.5 Flash: results close to flagship Seedance 2.5, with nearly twice the generation speed.
For 720p text-to-video and image-to-video generation, pricing is USD 0.152 per second, approximately 34% lower than Seedance 2.5.
The API will launch on BytePlusโs ModelArk platform on Sunday, November 15.
Watch the video below.
Eliana Skye ๐๐ retweeted
Drop a strawberry into a bowl. Now make it a rubber duck. Or change the liquid.
PAN 2, the world model the Institute of Foundation Models (IFM) just launched, simulates each version and shows how the splash changes.
IFM built PAN 2 to be grounded in time, physics, and space. The idea is that AI agents can check a what-if like this before they act. Below are three more demos, one for each.
๐ค Paid partnership ยท ๐ค Made with AI
This is a pretty cool idea:
Let your agent simulate each potential action before picking the best one.
The model is PAN 2, a world model from the Institute of Foundation Models (IFM). They call this "simulative reasoning".
The attached video shows an example with a surfer: same wave, three different moments to pop up. PAN 2 plays out all three, and only one of them catches the wave, so thatโs the one he goes with.
The model uses Generative Latent Prediction (GLP), and here is how it works:
โข In: the current scene plus an action, with how long to simulate
โข Out: what happens next, as a video you can check
There are two reasoning parts:
1. Conceptual reasoning: The LLM is the backbone here. It reasons about what happens next.
2. Perceptual reasoning: A 29B LDM fills in what it looks like.
This is all open under Apache 2.0!
They provided the weights, training code, and inference code, so you can run it and build on it yourself.
๐ค Paid partnership
Eliana Skye ๐๐ retweeted
MBZUAI's Institute of Foundation Models (IFM) just released PAN 2, an open foundation world model for physical intelligence.
It's built towards simulative reasoning: letting an agent play out what happens before it acts.
It works with two kinds of reasoning, each with its own job:
> Conceptual: an LLM works out what happens, when, and how the camera moves, then writes it down as text.
> Perceptual: a 29B diffusion model, trained from scratch, turns that text into what you see.
Take the LLM out, and multi-step tasks fall apart: 8.7% of steps right, no plan finished. Getting the order of actions right is the LLM's job.
With both, PAN 2 gets 75.1% of steps right and finishes 34.0% of plans in full, the highest among the models IFM tested. Every step has to hold for a plan to work.
More details in the technical blog:
ifm.ai/pan/
[1/1]
๐ค Paid partnership



















