GPT-Image-2 delivers some amazing results… and then there’s this. That’s one strong bond. 😭
English
daniel.dmai
272 posts
daniel.dmai
@danieldmai
Founder building scalable generative AI infrastructure
Sumali Aralık 2024
97 Sinusundan17 Mga Tagasunod
Bet you did not expect this happening. Watch until the end...
English
@flowersslop @mark_k And these results are impressive. I am just sad that these are no open source models
English
@mark_k in my tests none of the examples were more yellow than nb pro, for example this one
a lot of people are posting 1.5 images and think it is 2 bc some people say its already in chatgpt which it is not
x.com/flowersslop/st…
Flowers ☾@flowersslop
1 is nano banana pro 2 is maskingtape 3 is packing tape
English
@corbin_braun 1000% GPT Image 2 hooked with a agent and its done
English
@AIWarper i am not even hyped anymore.. i just wait to see more results and the 1 week aftermath, SD2 destroyed my excitement for new models
English
Can’t wait for GPT image 2 to drop
Everyone to say it’s SOTA and the best image generator ever made
Only for it to be nerfed into the ground and be completely useless a week later 🥲
Such is the OAI launch cycle
GIF
English
@NetworkChuck It's a constant switch from openai and anthropic at this point. Feels like a monthly back and forth.
but this made me think more about open source llms Qwen 3.6, Kimi, and even more.
Open Source needs to step up.
GIF
English
Anthropic says NO MORE OpenClaw, they are officially cutting us off starting April 4th.
How do you feel about this? Will this hurt Anthropic, OpenClaw or both?
English
@IamEmily2050 uff.. this rollout of Seedance models is a pure mess and still no real faces or issues even with ai faces. I love the quality but hate this
English
@aisearchio this slogan is burned into my brain every time i see new news thanks to you :D
GIF
English
Finally! Bytedance drops an image generator & editor that can run locally on your phone!
Creates 1024x1024 images in only 3 seconds
Will be open sourced soon!
carlofkl.github.io/dreamlite/
English
English
English
@mark_k finally someone said it. comparing a diffusion model with a autoregressive architecture model is always a pain for me.
English
If you can't tell apart diffusion from autoregressive image models, why are we even discussing? 😔
English
Umm… anyone wanna take a guess on what Elon means by this?
English
Never been able to explore latent space like this before..
👀
English
That i always get excited and then feel empty inside because its again.. not what i wanted to see.. :D
All I need:
Seedance 2.0 API
VEO 4
Wan 2.7
Mark Kretschmann@mark_k
Veo 4! ...would have been nice! 😉 But instead we're getting Veo 3.1 Lite from @GoogleDeepMind. A cheaper and faster variant of the established Veo 3.1. It's better than nothing and surely it's useful for some use-cases. You just have to wait a bit longer for Veo 4.
English