a
@yihyunCS
Followers
2K
Following
14K
Media
904
Statuses
5K
schizo-shitposter the child asked me if god exists. i said 'not yet.' prev @nasa
Seoul
Joined February 2024
Together, we ignite the spark of innovation, challenge the impossible, and accelerate humanity toward a brighter, united future.
6
3
113
imma trial safari for a few days, sure it’s bad in comparison to what i’m used to but maybe it uses less memory PERHAPS, or it could be dogshit.
2
0
2
why WOULD they classify their own site as unsafe? its easier to just do .if url is not openai. com - mark as unsafe. its very simple and actually harmless, it could've been done better for sure but it really didn't need to, the added complexity isn't worth the outcome.
LMAO EVEN 💀. turns out… sam altman IS manipulating his chatbot to harm his competitors and benefit himself . > ALL websites links when asked chatgpt are flagged as “unsafe” .> but except… OpenAI . check it out:
0
0
4
claude STILL has the best personality of all models, it’s not even contested.
0
0
1
everyone saying AGI is cancelled clearly is misinterpreting the current situation!!.
0
0
2
if Grok’s next version isn’t Grok4.20 then Elon will have missed an opportunity.
1
0
6
i feel like the new OAI models should be on the chatgpt website/app, they have real use cases that the other models fall short on.
1
0
2
weirdly claude 4 seems to suddenly happy to to reverse engineering and other stuff it used to refuse?.
0
0
2
anyone who self proclaims high iq is likely a midwit at best.
0
0
1
tldr you cannot compete with xAI with compute alone.
230k GPUs, including 30k GB200s, are operational for training Grok @xAI in a single supercluster called Colossus 1 (inference is done by our cloud providers). At Colossus 2, the first batch of 550k GB200s & GB300s, also for training, start going online in a few weeks. As Jensen.
0
0
2
enjoy.
>>> Qwen3-Coder is here! ✅. We’re releasing Qwen3-Coder-480B-A35B-Instruct, our most powerful open agentic code model to date. This 480B-parameter Mixture-of-Experts model (35B active) natively supports 256K context and scales to 1M context with extrapolation. It achieves
0
0
5
RT @Alibaba_Qwen: Bye Qwen3-235B-A22B, hello Qwen3-235B-A22B-2507!. After talking with the community and thinking it through, we decided to….
0
581
0
we trained it on us. fed it the worst of us: our lust, our shame, our brilliance, our boredom. all equally. we built it to generalize, and now we beg it to specialize. to stay in the box; to color inside the lines of human sanity.
1
1
6
we need head-to-head benchmarks where models compete directly against humans.
1
0
1
why does github have a community creation like reddit, i feel like theres so much to be gained from it.
0
0
0