Adrien Tsagliotis

Tech writer @JDNebusiness - Author @dunod - Lecturer @sorbonneparis1

[Adrien Tsagliotis @scoolada]

Wow - merci @joinstationf :) https://t.co/RBNMIjcq89

[STATION F @joinstationf]

Can’t be a good week end without a good book! Here are our favorites about startups and entrepreneurship 😉📚 @peterthiel @nireyal @scoolada @ericries @mikkelsvane @yegg @jwmares 🙏 #VendrediLecture #FridayFeeling

[Gensai.ai @ketsu]

Marc Theermann (Boston Dynamics) : "Boston Dynamics débutera la commercialisation du robot Atlas en début d'année", par @scoolada : https://t.co/ZO1NGZJSDy via @journaldunet

[Adrien Tsagliotis @scoolada]

@setu_ai_expert G

[Adrien Tsagliotis @scoolada]

@samruddhi_mokal Gamma

[Gilles Babinet @babgi]

La situation sur les multiples et la valorisation des principaux acteurs de l'IA fondationnelle.

[Ethan Mollick @emollick]

This paper shows that you can predict actual purchase intent (90% accuracy) by asking an LLM to impersonate a customer with a demographic profile, giving it a product & having it give its impressions, which another AI rates.

No fine-tuning or training & beats classic ML methods.

[Adrien Tsagliotis @scoolada]

@mikefutia Sora

[Adrien Tsagliotis @scoolada]

@eyad_khrais @perplexity_ai Prompt :)

[Adrien Tsagliotis @scoolada]

Interesting. I would bet on :
2 - 5 - 7 - 14 - 17

Let’s also see what Meta will release.

[GREG ISENBERG @gregisenberg]

i've been thinking about sora 2 (what it means for you, where the $$ can be made):

1/ someone will buy a synthetic Sora 2 style account for $10M and it will be a good deal but will look crazy

2/ whoever builds the “no-AI instagram” that takes off makes $100M-$1B. because social feeds will drown in Sora spam. networks that ban AI content, add verification layers will thrive.

3/ people start selling their own likeness. “rent my face for $99/month.” agencies become gatekeepers of likeness rights. the scarjo fight is the first domino.

4/ someone will build a daily Sora channel, flip it into a $50M+ AI startup within 18 months completely bootstrapped.

5/ in 3 months, feeds will be unrecognizable. 1-person “ai-first creators” dropping 200 vids/day. the ones who figure out hooks + branding → next gen MrBeasts.

6/ creation goes to zero cost, so ppl who own the pipes (big accounts, media pages, curation brands) will tax the flood. 2010s meme pages all over again, but cinema-quality.

7/ someone will mint a Sora-native character that builds an empire (music, merch, movies). think hello kitty x mrbeast. content making influencers are the new influencers

8/ people start selling their own likeness. “rent my face for $99/month.” agencies become gatekeepers of likeness rights. the scarjo fight is the first domino.

9/ 1,000 ad variations overnight is the new normal → cpm wars. who wins? the ppl who build the distribution machines that can test + filter that flood.

10/ tooling gold rush - watermarking, copyright detection, brand overlays. when the world drowns in video, verification tools become choke points.

11/ movies won’t die, but they’ll be swallowed by endless serialized micro-shows. one breakout sora-native series hits, netflix panics.

12/ memes/genres that used to last years now die in weeks. whoever masters the rhythm of fast-burn cycles builds audience the fastest.

13/ platforms double dip - they’ll host the flood and sell you boosts to cut through it. organic reach might fall (i hope not, but manage your risk accordingly)

14/ vertical content studios rise - apps for niches: Sora-for-fitness, Sora-for-real estate listings, Sora-for-crypto memes. packaged prompts + workflows + export buttons.

15/ hollywood talent agencies will pivot into “likeness funds.” they’ll securitize the rights to actors’ faces and voices like IP portfolios. sidenote: hollywood prob not feeling so great about this

16/ Synthetic “how-to” teachers will dominate niches. Think Khan Academy, but cloned 1,000x and personalized to every learner.

17/ distribution curators will be the new venture capitalists. Whoever can guarantee views will become gatekeepers.

18/ copyright courts will get nuked. The first $1B+ lawsuit over Sora likeness rights will set precedent for a new industry.

19/ a religion will literally form around a Sora-generated prophet. sounds insane. it will happen.

20/ synthetic nostalgia startups will raise hundreds of millions. recreate your 7th birthday, your grandma’s kitchen, your lost dog. It’ll be addictive.

21/ a real estate startup will generate cinematic walk-throughs of homes before they’re built, then sell them instantly. $1B company.

22/ in 5-10 years, people won’t ask “what’s your favorite show?” they’ll ask “what’s your favorite generator?”

in the next tweet I’ll outline the steps I’d personally take to make $$/earn distribution from sora 2 in a few steps

sora 2 is a big deal

i can't sleep thinking about it, idk about you

[Adrien Tsagliotis @scoolada]

@shl 🙋‍♂️

[Adrien Tsagliotis @scoolada]

😯

[OpenAI @OpenAI]

Sound on.

[Andrew Ng @AndrewYNg]

The invention of modern writing instruments like the typewriter made writing easier, but they also led to the rise of writer’s block, where deciding what to write became the bottleneck. Similarly, the invention of agentic coding assistants has led to a new builder’s block, where the holdup is deciding what to build. I call this the Product Management Bottleneck.

Product management is the art and science of deciding what to build. Because highly agentic coding accelerates the writing of software to a given product specification, deciding what to build is the new bottleneck, especially in early-stage projects. As the teams I work with take advantage of agentic coders, I increasingly value product managers (PMs) who have very high user empathy and can make product decisions quickly, so the speed of product decision-making matches the speed of coding.

PMs with high user empathy can make decisions by gut and get them right a lot of the time. As new information comes in, they can keep refining their mental models of what users like or do not like — and thereby refine their gut — and keep making fast decisions of increasing quality.

Many tactics are available to get user feedback and other forms of data that shape our beliefs about users. They include conversations with a handful of users, focus groups, surveys, and A/B tests on scaled products. But to drive progress at GenAI speed, I find that synthesizing all these sources of data in a PM's gut helps us move faster.

Let me illustrate with an example. Recently, my team debated which of 4 features users would prefer. I had my instincts, but none of us were sure, so we surveyed about 1,000 users. The results contradicted my initial beliefs — I was wrong! So what was the right thing to do at this point?
- Option 1: Go by the survey and build what users told us clearly they prefer.
- Option 2: Examine the survey data in detail to see how it changes my beliefs about what users want. That is, refine my mental model of users. Then use my revised mental model to decide what to do.

Even though some would consider Option 1 the “data-driven” way to make decisions, I consider this an inferior approach for most projects. Surveys may be flawed. Further, taking time to run a survey before making a decision results in slow decision-making.

In contrast, using Option 2, the survey results give much more generalizable information that can help me shape not just this decision, but many others as well. And it lets me process this one piece of data alongside all the user conversations, surveys, market reports, and observations of user behavior when they’re engaging with our product to form a much fuller view on how to serve users. Ultimately, that mental model drives my product decisions.

Of course, this technique does not always scale. For example, with programmatic online advertising in which AI might try to optimize the number of clicks on ads shown, an automated system conducts far more experiments in parallel and gathers data on what users do and do not click on, to filter through a PM's mental model of users. When a system needs to make a huge number of decisions, such as what ads to show (or products to recommend) on a huge number of pages, PM review and human intuition do not scale.

But in products where a team is making a small number of critical decisions such as what key features to prioritize, I find that data — used to help build a good mental model of the user, which is then applied to make decisions very quickly — is still the best way to drive rapid progress and relieve the Product Management Bottleneck.

[Original text: https://t.co/1tulDs3k7U ]

[Adrien Tsagliotis @scoolada]

Very interesting demo on how to make infinite videos

[fabian @fabianstelzer]

Kling 2.5 "infinite" videos are HYPE rn, so here's a 15 min full tutorial that shows you exactly how you can create your own - I made a custom agent that does everything for you so you can focus on the creative direction:

[Grant Lee @thisisgrantlee]

Yes, our team of 30 could fit in a small restaurant. But we serve 50 million users profitably while most AI startups burn millions serving thousands.

The future belongs to tiny teams of extraordinary people.

[Travis Davids @MrDavids1]

I think I'm the first person to generate more than 8 items into a single image using Google Gemini Flash Image (Nano Banana). I have even exceeded the 8 upload limit on Freepik. How did I do this? Create a collage with everything and label each item on the image. When you upload the image, be descriptive and give each item the same name you labeled it on the image. Look at the end result, 10 items in a single image with excellent accuracy (Only thing not super accurate is the watch). I think this method is even more accurate than uploading individual photos. Go and try it!

The prompt I used to generate 10 items into a single image is as follows: A man is standing in a modern electronic store analyzing a digital camera. He is wearing a watch. On the table in front of him are sunglasses, headphones on a stand, a shoe, a helmet and a sneaker, a white sneaker and a black sneaker

[Adrien Tsagliotis @scoolada]

@framer_x I actually used Einstein as a character for consistency.

Script (and « jokes ») by ChatGPT, voice with ElevenLabs

[Adrien Tsagliotis @scoolada]

@framer_x Cool! These AI models enabling multi-shot generation are such a game changer and a huge time saver.

I made this quick video back in January :

At that time everything was done with Runway and Kling, without any multi-shot option (which would’ve made the video more dynamic).

[Adrien Tsagliotis @scoolada]

@SOL_Trending @framer_x Yes, it’s because I actually screenshotted the CapCut result

Nope, not available on YouTube (it was just a quick video I made for my nephews and nieces and to try these multi-shots models)

[Adrien Tsagliotis @scoolada]

The next Walt Disney could be a 12-year-old making a brilliant movie from their bedroom and sharing it on YouTube. These are exciting times! 🍿

[Framer 🇱🇹 @Framer_X]

This is Hopeless Steve.

A cartoon character you've created with AI.

You grew up watching Simpsons, South Park and now you dream of Hopeless Steve becoming the next big IP.

So you start posting short sketches on Youtube, TikTok, IG, etc.

Then one day... 👇

[Adrien Tsagliotis @scoolada]

@liboar @framer_x As mentioned, this was 100% AI-generated as part of a tech test to see what these tools could do

And yes, my nephews and nieces find it fun (and will still happily watch Pixar, which thankfully creates way cooler and more creative content than this little video 😉)

[Adrien Tsagliotis @scoolada]

@framer_x I usually ask ChatGPT to keep the same character and place them in a different situation, since Veo3 isn’t always great at generating a totally unrelated scene.

What tool do you use for switching from one scene to another?

[Adrien Tsagliotis @scoolada]

That’s very cool! 🎥

[Stelfie the Time Traveller @StelfieTT]

This was a complicated one.

My latest fully AI short movie : "One Way"

created thanks to
@runwayml (Aleph), @HeyGen_Official , Google Veo.