If the next @XBOX support physical disc it's over for Sony for the next generation. No portable console gimmick will save them. The ball is in your hand @Microsoft.
On #gtavıleak People are overestimating the impact this has. In the end they will sell as many copies. The only one losing are the fans who got a low quality gameplay first impression and the dummies who bought the scam crypto coin.
What kind of lunatic work at @Reddit ? They send email about potential compromised account asking to reset your password with links inside the email that doesn't link to reddit.com but redditmail.com . Every fishing flags are in there.
They all do it btw. Disgusting practice.
CLAUDE OPUS 4.6 IS NERFED.
BridgeBench just proved it.
Last week Claude Opus 4.6 ranked #2 on the Hallucination benchmark with an accuracy of 83.3%.
Today Claude Opus 4.6 was retested and it fell to #10 on the leaderboard with an accuracy of only 68.3%.
A 98% increase in hallucination.
bridgebench.ai just confirmed that Claude Opus 4.6 has reduced reasoning levels and is nerfed.
Readers added context they thought people might want to know
The cited BridgeBench scores for Claude Opus 4.6 are from tests with different scopes: 6 tasks initially (83.3% accuracy) vs. 30 tasks recently (68.3% accuracy). Analysis of the 6 common tasks shows similar performance (87.6% vs. 85.4%), suggesting no major change.
bridgebench.ai/hallucination/…
bridgebench.ai/hallucination/…
x.com/paul_cal/statu…
HeyLGros retweeted
Seedance 2.0 landing just in time for HappyHorse to take it out behind the shed and put it to bed.
This will go down in history as one of the biggest fumbles.
They completely nuked their momentum
Today we are rolling out Dreamina Seedance 2.0 in the U.S.
And your first try is on us:
→ Everyone will have one free trial of Dreamina Seedance 2.0 across CapCut’s app, desktop and web.
→ New users get 90% OFF their first month of CapCut Pro (limited time)
Imagination has no ceiling now. Enjoy creating non-stop.
Access Dreamina Seedance 2.0 in CapCut:
→ AI Lab (app, v17.1.0)
→ Video Studio (web, capcut.com/ai-creator-home)
→ AI Video (app and desktop)
It might take some time for the model to reach everyone.
As the model rolls out, we're partnering with experts and creative communities to learn, iterate, and continuously improve on the ways this model can empower human creativity. Learn more here: capcut.com/newsroom/dreamina…
RT+Comment in 4hr to get extra 1000 credit in your DM!
In our business we generate around $1M worth of images through a specific API. This week I had the possibility to move our service to WAN 2.7 image which fits our requirements. With what happed with Qwen image 2 and WAN 2.5, I decided to not do it. Feels good.🖕@Alibaba_Wan .
While @Alibaba_Qwen is locking their recent 3.6 weight and is switching to a pure closed business model, Google delivers to the Open Weight community. Well done.
HeyLGros retweeted
Dana White reacts to fans complaining about the UFC using AI for promo videos:
"Give me a f*cking break. AI is coming, and if we're using AI who gives a sh*t. People are upset, and we should use artists?
How about this. Shut the f*ck up and watch the fights."
HeyLGros retweeted
Calibri.
Ultra-lightweight post-hoc calibration for DiTs.
- 2x speedup (FLUX 30 -> 15 steps).
- 3x for Qwen (100 -> 30).
- Zero extra compute or params;
- Enhances text-prompt alignment, visual fidelity;
- No memory and inference overhead.
v-gen-ai.github.io/Calibri-p…
After further testing of @PhotaLabs' Image "Model", we have concluded that Phota is merely a Nano Banana Wrapper that is being sold as a 'new foundation model'.
This is deceptive and doesn't align with our values, nor our value proposition.
We have therefore decided to withdraw this model again, after having added it earlier today.
We apologize for any confusion caused, and ask PhotaLabs to adjust their communication.
HeyLGros retweeted
PyTorch 2.11 is here!
> Distributed training and attention optimized;
> FlexAttention hits 3x speedups on Hopper/Blackwell via FlashAttention-4;
> backprop through collectives;
> CUDA 13 becomes default;
> TorchScript deprecated
Nice!
pytorch.org/blog/pytorch-2-1…
Bytedance should release the weight to the open source community. There will be so much copyrighted content everywhere that Hollywood will give up on their witch hunt and Seedream will stop being the target.
I just read this Reuters article about Seedance 2.0. Not sure if it’s fully confirmed or still speculation, but the content of the article feels very similar to the previous news when Disney raised copyright concerns about AI models. If it actually turns out that Seedance 2.0 doesn’t get released, that would be pretty disappointing considering how promising the tech looks for creators. Hopefully it’s just a delay while they sort things out behind the scenes. what do you guys think if it really ends up not being released?
link in the comments below. 👇
