MacWhisper is good for that too.
GENERATIVE AI
-

AI models struggle with sensitive topics and prompt limitations
By
–
Y no es sólo con ese prompt concreto, sino con según que temas delicados le preguntes.
-
Grok 4 Self-Sabotaging With Elon Musk Opinion Searches
By
–
Es de coña y me resulta increíble lo de Grok 4 haciendo búsquedas a las opiniones de Elon Musk para responder cualquier tema controvertido que le plantees.
— Carlos Santana (@DotCSV) 11 juillet 2025
Acabo de hacer la prueba y efectivamente sucede, qué manera más tonta de auto-sabotear a la IA que iba a buscar la "verdad" pic.twitter.com/JY9lxjPUCNEs de coña y me resulta increíble lo de Grok 4 haciendo búsquedas a las opiniones de Elon Musk para responder cualquier tema controvertido que le plantees. Acabo de hacer la prueba y efectivamente sucede, qué manera más tonta de auto-sabotear a la IA que iba a buscar la "verdad"
-

Hailuo 2 AI Video Generation Test: Barrel Roll Physics
By
–
Barrel roll test. For a moment I thought Hailuo 2 had managed this one.
— fofr (@fofrAI) 11 juillet 2025
> A sleek sports car launches off a curved ramp and performs multiple barrel rolls through the air while spinning end-over-end, with its wheels rotating at different speeds as it tumbles gracefully against a… https://t.co/NjqomJ2r6U pic.twitter.com/QWwOBrdH72Barrel roll test. For a moment I thought Hailuo 2 had managed this one. > A sleek sports car launches off a curved ramp and performs multiple barrel rolls through the air while spinning end-over-end, with its wheels rotating at different speeds as it tumbles gracefully against a
-

Grok 4 Relies on Web Search Results and Online Code
By
–
Grok 4, in general, is very influenced by search results and pretty credulous when it sees a web search result. When you ask it to code, it often looks for code online first and uses that.
-
Grok 4: Intelligence or Strategic Positioning?
By
–
Everyone’s talking about Grok 4.
Charts. Benchmarks. Elon. Hype.
But here's the real question:
Is Grok 4 the smartest AI ever built — or just the most strategically positioned? -
Replicate Corrects Veo 3 Fast Model Pricing Error
By
–
Sorry folks, we got the pricing on this one wrong at launch.
— Replicate (@replicate) 11 juillet 2025
The Veo 3 fast model is $0.40/s, which is $3.20 per video. This is about half the price of Veo 3. It is not 5x cheaper. https://t.co/O2UH9gSm4mSorry folks, we got the pricing on this one wrong at launch. The Veo 3 fast model is $0.40/s, which is $3.20 per video. This is about half the price of Veo 3. It is not 5x cheaper.
-
Replicate Corrects Pricing Error for Video Generation Model
By
–
Sorry folks, we got the pricing on this one wrong at launch. The model is $0.40/s, which is $3.20 per video. This is almost half the price of Veo 3. It is not 5x cheaper.
-

GPT-5 and Grok 4 Race Toward ASI Singularity Acceleration
By
–
My friend @iampiet nailed it with this graphic showing where we're at today. Yes, Grok 4 is great but GPT-5 will come soon and probably be even better. And the cycle continues. This acceleration is not going to stop. We're in the ASI singularity.
-

Grok 4 Heavy: ASI-Level Capabilities and Rapid AI Progress
By
–
@beffjezos thinks Grok 4 Heavy is already ASI level. I believe Grok 4 + better tool usage is already what most people would have considered AGI 10 years ago. Reminder: Grok 3 came out in February of this year… the rate of progress is insane.