Clement Tan

Senior Creative Designer | Prev: Razer, Secretlab | No AI

[Clement Tan @xavierclement77]

My final project is finally released… Many thanks to those who’ve helped and trusted my art direction and design. FFXIV has a special place in my heart, and it’s still the best game for me. It’s a little poetic and bittersweet to close the chapter with this collab 🥹 #ffxiv

[Secretlab @secretlabchairs]

Traverse Eorzea in comfort with the official Secretlab TITAN Evo FINAL FANTASY XIV Edition, designed in collaboration with the FINAL FANTASY XIV team: https://t.co/VoXuhcoiKb The ultimate backup for every adventurer, and a steadfast companion to support you on your journey.

[Ed Newton-Rex @ednewtonrex]

Google is claiming it has the right to train gen AI on people’s YouTube videos, because of terms users signed that don’t mention AI and that no one reads anyway.

I suspect that if people aren’t angry about this, it’s because they don’t know about it.

https://t.co/IrlmgEZAmj

[Karla Ortiz @kortizart]

One more note on this. Epic utilizes Google’s NanoBanana.

Google is currently being sued by artists due to their exploitative use of datasets like LAION— which are littered with ill-gotten copyrighted works— to train the models that powers this nonsense.

[Fortune K. @Misfortuneee]

Experimentation doesn't require AI tools. If the desire is to making things people have seen a million times already and to burn down a forest then AI is the perfect tool.

[Ed Newton-Rex @ednewtonrex]

This is disgusting behaviour from Google.

It is using its monopoly power to try to force creatives to let it train on their work.

It is career suicide for artists not to upload to YouTube. Arguing that doing so allows Google to train AI on their videos/music - AI that competes with them - is outrageous. Whatever the Ts & Cs say.

[Ed Newton-Rex @ednewtonrex]

These two planes flew over an event where the CEO of AI music company Suno was speaking this week

https://t.co/aTPLzgWCzR

[Ed Newton-Rex @ednewtonrex]

Claude 1 was trained on pirated books.
Claude 2 was trained on pirated books.
GPT-3 was trained on pirated books.

The AI industry was built on piracy.

[Esteban Rafael • Tech IA @IAbyEsteban]

⚠️ La parte más honesta de toda la industria de la IA probablemente se filtró por accidente.

Mientras algunos CEOs admiten que harán falta entre 6 y 8 billones de dólares en infraestructura para justificar una demanda que todavía no existe, ahora aparece un documento interno de Microsoft donde la primera fase de su nuevo asistente Scout sería: “hacer a la gente adicta”.

Muchos se enfocarán en la palabra. Yo me enfoco en el modelo.

Porque cuando una tecnología necesita cantidades históricas de capital, centros de datos gigantescos, consumo energético masivo y márgenes que aún no aparecen, el activo más valioso deja de ser la innovación y pasa a ser la dependencia.

Las redes sociales ya demostraron que la atención humana puede monetizarse mejor que cualquier producto físico. Ahora la siguiente fase parece ser convertir la IA en una extensión permanente de la mente, del trabajo y de la vida diaria.

- No quieren que uses la IA.

- Quieren que no puedas funcionar sin ella.

Y cuando una empresa diseña una tecnología para generar dependencia antes que valor, deja de construir herramientas y empieza a construir hábitos.

Las redes sociales nos vendieron conexión y terminaron multiplicando problemas de ansiedad, aislamiento y adicción digital. La pregunta es si estamos viendo el mismo experimento, pero esta vez a una escala mucho mayor.

[Ed Newton-Rex @ednewtonrex]

Suno - a company that trained on "essentially all music files of reasonable quality that are accessible on the open Internet", and argues it does not need to pay to do so - is now valued at $5.4 billion.

They will pay for GPUs, and for engineers, but not, apparently, for the music they use - other people's music - without which their models would not work.

I know lots of musicians who will be upset. I am with them.

[Karla Ortiz @kortizart]

“The thief is here. We have to observe it. We have to lean into” I spinelessly say as the thief steals all my stuff.

[Reid Southen @Rahll]

Remember when ChatGPT helped someone plan a mass shooting, and a dozen OpenAI employees notified management, and they did nothing and 8 people died?

They don't want to be liable for their incompetence and role in such things, despite doing absolutely nothing to prevent it.

[BuBBliK @k1rallik]

🚨 Maine is about to become the first state in US history to BAN data center construction

AI needs servers. Servers need land, water, electricity.
And Americans are starting to say: not in my backyard.

Microsoft, Google, Amazon are spending hundreds of billions on AI infrastructure RIGHT NOW.

The first wall just appeared.

Watch what happens when other states follow.

[Reid Southen @Rahll]

"This is the worst it will ever be."

[Reid Southen @Rahll]

Aside from the energy costs, the regulation OpenAI is having trouble with is in fact copyright law.

[🏮Zakuga Art🏮 @ZakugaMignon]

I heard Deviantart is struggling, and desperately asking for people to start a core account & come back. lol

Good die with adobe you traitor, I hope your company files for bankruptcy. Making a genAI out of the art we posted for years in support of your site deserves bankruptcy

[Joseph O'Neill @JosephONeillx]

Quite simply, this must be stopped.

[Reid Southen @Rahll]

You don't hate AI enough.

[Bryce Greene @TheGreeneBJ]

KEEP FIGHTING DATA CENTERS

[Karla Ortiz @kortizart]

This man is not only a lunatic but also mobster, and every coward in congress who refuses to act is equally complicit.

GET HIM THE HELL OUT OF THERE NOW!

[NIK @ns123abc]

🚨BREAKING: OpenAI shares have fallen out of favor and investors are trying to sell but finding no buyers

Sam Altman 4 months ago:
“if you want to sell your shares, I'll find you a buyer. Enough! I think there's a lot of people who would love to buy OpenAI shares. We can sell your shares very quickly!”

LMFAOOOO this aged so badly

[Simplifying AI @simplifyinAI]

🚨 BREAKING: OpenAI and Google are about to have a massive legal problem.

OpenAI, Google, and Anthropic have repeatedly sworn to courts that their models do not store exact copies of copyrighted books.

They claim their "safety training" prevents regurgitation.

Researchers just dropped a paper called "Alignment Whack-a-Mole" that proves otherwise.

They didn't use complex jailbreaks or malicious prompts.

They just took GPT-4o, Gemini, and DeepSeek, and fine-tuned them on a normal, benign task: expanding plot summaries into full text.

The safety guardrails instantly collapsed.

Without ever seeing the actual book text in the prompt, the models started spitting out exact, verbatim copies of copyrighted books.

Up to 90% of entire novels, word-for-word. Continuous passages exceeding 460 words at a time.

But here is the part that changes everything.

They fine-tuned a model exclusively on Haruki Murakami novels.

It didn't just learn Murakami. It unlocked the verbatim text of over 30 completely unrelated authors across different genres.

The AI wasn't learning the text during fine-tuning.

The text was already permanently trapped inside its weights from pre-training. The fine-tuning just turned off the filter.

It gets worse.

They tested models from three completely different tech giants. All three had memorized the exact same books, in the exact same spots.

A 90% overlap. It's a fundamental, industry-wide vulnerability.

For years, AI companies have argued in court that their models are just "learning patterns," not storing raw data.

This paper provides the smoking gun.