
CEO & Founder @Magnific (formerly Freepik) • Ex @Google • Co-founder @BeSoccer_es (acq Livesport) • Founder Panoramio (acq by Google)
London, England
Joined September 2008
- Tweets5K
- Following579
- Followers12.4K
- Likes2.5K
Pinned Tweet
This is the biggest announcement I've made in the history of the company
In 2010 we started Freepik. Today, what we built has outgrown that name.
It's time we change that. We are Magnific
Nvidia live: Jensen Huang resigns as chief executive after affair allegations ft.com/content/b4a0f3d9-f075… via @ft
4
5
1,147
Joaquín Cuenca Abela retweeted
Chapter Seven of Season One of the Magnific Original Series The Chronicles of Bone is HERE! This is only the first 15 minutes of the Episode - to see the WHOLE episode, see the comment section! 😎
Today is also a very special day, because The Chronicles of Bone officially turns ONE YEAR OLD! 🥳
I still remember pitching The Chronicles of Bone to studios and potential partners, only to hear the same response: “I don’t get it.” But I believed in this world, so I stopped waiting for permission and started bringing it to life myself.
Then came Ángela and the incredible team at @magnific They saw the vision, believed in me, and took a chance on this crazy idea. And just like that, The Chronicles of Bone became a Magnific Original Series.
One year later, we’re averaging millions of views per episode, and my YouTube channel has grown from 18,000 subscribers to over 1.4 MILLION! It’s honestly surreal.
But the greatest part isn’t the numbers. It’s knowing that something I dreamed up has connected with so many people around the world.
To Magnific and every single one of you who watched, commented, shared, and believed in this story when so many didn’t: THANK YOU. I’ll never have the words to express what your support means to me.
One year ago, The Chronicles of Bone was just a dream. Today, it’s a world we all share. And we’re only getting started!
Created entirely with the tools inside Magnific - Chapter Seven
Stranded on Dead Man’s Bay in the aftermath of the Battle of the Bay, Hook and Anne find themselves with no ship, no crew, and no way off the island. Desperate for escape, Hook devises one final gamble: a journey to the neighboring Cannibal’s Cove, ruled by the ruthless and unpredictable Mad Pirate King Crowe. Meanwhile, in Highlorn, Varek reunites with his loyal Shadow Knight and sets a deadly plan into motion to eliminate the notorious outlaw Robin Hood and his crew once and for all.
#thechroniclesofbone #cob #magnificoriginals
what a great talk by @lil_dill youtu.be/GLvFTMtw4Jk?si=hwGI…
it even uses examples from Gutenberg's bible, my long time favorite example of craft. A must watch
2
25
4,152
Joaquín Cuenca Abela retweeted
Bringing greater creative control to AI-generated videos.
At #ApsaraConference2026 in Amsterdam, Joaquín Cuenca Abela @cuenca, CEO & Co-Founder of @magnific, highlighted how Wan's multimodal references enable greater consistency across characters, locations and shots, opening up new possibilities for AI-generated video.
Explore more: click.alibabacloud.com/m/200…
#ApsaraAmsterdam #AgenticEra #AIAgents #AlibabaCloud
Hace muchos años hablando con un colega en Google sobre la seguridad de los sistemas (de software) modernos. En teoría hay 10 capas de seguridad y tienes que encadenar muchos fallos de seguridad para conseguir un ataque. En la práctica son todas una mierda, nadie las ha probado de verdad en años porque ninguna tiene “estrés”, ningún sistema depende al 100% de una sola de ellas y eso hace que no se miren con lupa. Al final dependes de 1.5 capas, se hace funcionar la cosa y aguanta un poco caídas, pero si 2 cosas caen al mismo tiempo estás ya en terra incógnita y las demás capas no aguantan un soplo.
Los audios de las conversaciones del Centro de Control señalan otro FALLO: no parecen conscientes de estar gestionando una emergencia.
Esa es mi impresión.
Los sonidos son combinación de desorden y calma injustificada.
Hay órdenes erróneas. Pero es peor que eso: el procedimiento sorprende por manual y humano. Me puedo equivocar. Pero es chocante.
Al caer el LZB, y con trenes que (parece) todavía no han alcanzado ninguna señal lateral que los proteja, uno pensaría en máxima precaución, garantías, baja velocidad. Pero no se trasluce eso ni de los audios ni de los hechos conocidos.
¿Pueden faltarnos claves? Sí.
Pero la sensación ahora mismo es que falla el LZB, y fallan las órdenes, y falla el sistema en su conjunto cuando pone a tres personas con un manual en un “punto único de fallo”.
También me recuerda el incidente reciente de Claude en el que por error lanzaron a producción una versión de Code con el código fuente. El post mortem del lider de Claude Code era que la solución era automatizar MÁS. Me gustaría saber cómo de normal es que se caiga el LZB? Cuántas veces ha caído este año? Por el tono de voz en las conversaciones parece algo totalmente rutinario
3
8
1,574
AI models are the new CPUs, and we’re building the OS. Insane new functionality in your favorite platform 💜
me uno a los que reescriben autofirma. Aquí tenéis una versión en línea de comandos. Especialmente útil si lo usas como skill en Claude / Chatgpt, solo le tienes que dar el pdf y le pides que te lo firme. Encuentra dónde poner la firma visible automáticamente github.com/e98cuenc/firma-cl…
Joaquín Cuenca Abela retweeted
The biggest risk for a successful company is believing that what made it successful will keep working forever.
It took Freepik 12 years to grow from scratch into a large, successful business.
It was a remarkable journey. We had millions of users and were competing with the biggest players in the stock content industry.
Around 4 years ago, when Dall-e 2 was released, I was with @cuenca at a founders event in Malaga and he told me he couldn’t sleep for days thinking about how this was going to change our industry. It was going to redefine how creative content is produced, distributed, and consumed.
Many in our user community hated AI-generated content. Our competitors were suing AI model providers. And our business was doing incredibly well.
Still, we went all in on AI. We decided to transform Freepik into an AI-native company.
Freepik became @magnific. And in less than three years, we grew the business 10x.
Shutterstock, once one of the giants of stock content, has seen its market capitalization collapse from over $4 billion in 2021 to approximately $170 million today. Getty Images, another iconic name in our industry, filed for Chapter 11 bankruptcy protection last week.
For decades, the creative industry was built around distributing existing content. Now, increasingly, it’s about generating entirely new content. That represents a fundamental shift in where value is created.
And when that happens, being an established market leader can become a disadvantage if you’re too attached to the business model that got you there.
We didn’t reinvent Freepik because it was failing. We reinvented it precisely because it was successful, and we wanted to make sure it stayed that way. Today, Magnific is a very different company from the one we were three years ago. And I suspect we’ll be a very different company three years from now.
Your past success is an asset. Until you start treating it as something you need to protect.
Chumbo started with concept art, before there was even a story
@Cont_animation combined character design and art direction with Seedance 2.5 animation on Magnific, then shaped the film through hands-on compositing, human voice performances and original music
Learn how it was made in the thread
Too much reality?
We've got just the thing. 🧪
CHUMBO, issue 2 of the Contanimation TV anthology, a @magnific Original by Contanimation.
Otro trabajazo de @Cont_animation 💜💉💉💉
Too much reality?
We've got just the thing. 🧪
CHUMBO, issue 2 of the Contanimation TV anthology, a @magnific Original by Contanimation.
you don't know how difficult it is to get this from Ana :)
Did @magnific really just release a model that's better than @midjourney?!? Seriously impressive! Congrats!
🇪🇸🫶🇪🇺💪
2
19
2,240
Today's AI models are a technical marvel. With enough expertise, you can steer them almost anywhere. But by default they have to serve many masters, and easy prompts give you bland results.
Magnific is not build for "anyone." We serve creatives, and the default mode on these models is not what our users want.
Months ago we started a project with these goals:
1) built on top of existing models. Leverage the existing AI models as the building block. Get better when models get better.
2) make the default generation by and for creatives. Easy to get great results.
3) trivial to stay on brand.
A new model from scratch will be something that will become obsolete in weeks / months. Our approach enhances them for our users, making a useful contribution that is long-lasting and differentiated.
After many false starts, we landed on a pre, post-processing, and routing layer on top of a few of today's state of the art models.
We tested the model first with our employees, and later with external users, on two tests: best image for a given prompt, and best image overall (blinding the prompt). The first checks prompt adherence, the last checks image aesthetics.
Magnific One came out first on both tests by a wide margin.
You can use it exclusively on @magnific, test it yourself and tells us what you think.













