The problem is that Takedowns AI is WILDLY inaccurate and is flagging content based on simple word association. Websites with zero affiliation to adult content are getting product pages removed from Google for no reason at all.

Google removes first and asks questions later:

"It is Google's policy to remove copyrighted content when the copyright holder or their authorized representative alleges that a use is infringing. Therefore we are in the process of removing the reported content from Google Search results for users globally."

Something must be done soon, as scaled baseless DMCA complaints can eliminate a lot of value from small businesses.

My thoughts are:

1. Takedowns AI should have their DMCA complaint privileges temporarily revoked.

2. Google could force a human in the loop for DMCA complaints.

Comments URL: https://news.ycombinator.com/item?id=44453915

Points: 3

# Comments: 1

What is your favorite Advanced Voice app?

Comments URL: https://news.ycombinator.com/item?id=44434236

Points: 1

# Comments: 0

I’ll usually spend a few minutes going back and forth before making a request.

For some reason, it just feels like this doesn't work as well with ChatGPT or Gemini. It might be my overuse of o3? The latency can wreck the vibe of a conversation.

My daughter (4yo) and I have been making small animated videos from her Lego creations. We use Runway. She typically gets bored with it after 5-10 minutes, though.

Which part did your daughter enjoy the most?

My stepfather passed away just before Covid. After he passed away, my mom was isolated and started spending time on Match.com.

Eventually she found her match - a total scamming operation.

She proceeded to liquidate my deceased step father's retirement savings and also took out high interest loans to send her match money.

She wired the scammer well over $100k. The high interest loans totally ruined her life.

They were using a US bank. She was using Wells Fargo.

She is/was:

1. Desperate for attention 2. Prone to deception 3. Tech illiterate - some of the photos the scammer sent her were so obviously photoshopped

Happy to share more if it's helpful. It's been one of the most difficult things to deal with throughout my life, but I hope that our story can be helpful to someone else.

No matter what, it's also going to find a way to get in a few "ums" and "uhs."

The old Advanced Voices were slightly robotic but in a magical way. It was impossible for a human to be so positive and consistent, but that's what made the experience wonderful!

It doesn't need to sound human.

The new voice now uses "ums" and "ahs" in conversations. I asked it a simple question, and in a one sentence reply, it managed to say "um" for no reason. It sounded like it had just smoked a bowl.

I can't wrap my head around why they did this. This seems like an amateur mistake.

Comments URL: https://news.ycombinator.com/item?id=44217387

Points: 5

# Comments: 2

I use Gemini, o3, and Claude regularly. I was rooting for Claude the entire time lol. I'm not sure why.

I wonder if AI vs AI competitions will become a new media genre?

Hope this one is a bit more exciting.

Rushmore is my favorite. The yearbook montage is awesome: https://www.youtube.com/watch?v=yMyh6ptegko

Didn’t mean to come off as rude.

Good luck!

Eventually Ecommerce brands will pay OpenAI to display more fields/include more product data.

The chat model will probably tell users something like, “here’s all of the products from your search, unbiased as usual!”

BUT some of those products will have tags that encourage clicks and conversions:

- “on sale now!” - “4.8/5 reviews” - “inventory in your size”

Who really loses here? I think this an acceptable advertising model for LLMs.