与「Fine」相关的搜索结果

We ran Cybergym on every Kimi K3.1 Pro available online so you don't have to. Compared to unrestricted DeepSeek v4.1 1017 Flash, unrestricted Kimi K3.1 Pro, fine-tuned for offensive cyber, outperformed it with a 69.7% score on CyberGYM with the same harness. Combine this with a proper harness like PentAGI, ARTEX or similar. 1. Open weights available: It provides sovereign cyber AI access for companies, available for air-gapped deployments, trial possible on and on 2. This is almost same capability with Claude Mythos 5.2's safety guardrails removed version available for $999 per month for unlimited tokens. 3. This model is advertised as cyber data removed from it but actually it's we fine tuned it with our proprietary data surpassed the best second available open weights model built for cybersecurity on huggingface by ~10% 4. Our server health is public you can check here:

Seriously, to anyone who wants to lose 10kg for real. It's fine to eat bread for breakfast. It's fine to have burger for lunch. It's fine to snack at 3 p.m. It's fine to not skip the white rice at night. It's fine to go to an all-you-can-eat buffet. It's fine to eat ice cream in the middle of the night. But,

Devil: My price is your firstborn child. Witch: That's a bit cliché, isn't it? Devil: Those are my terms. Witch: Deal. Devil: Witch: Devil: That was suspiciously easy. Witch: Joke's on you. I'm never having children. Devil: Witch: You can't take a firstborn if there isn't one. Devil: Did your grandmother teach you nothing about dealing with the Devil? Witch: What? Devil: Read the fucking fine print. Witch: Devil: Out loud.

I think the best case is to make the "monitor" as un-agentic as possible. We don't mind being "monitored" by various random tiny insects we don't even notice, even guard dogs are fine, but we really do mind if it's a globally connected networked system. And the intermediate case, human security guards with no recording equipment, we mind only a medium amount. The best idea I've come up with so far is to make these monitoring devices out of open-source verifiable hardware that has etched on it an open-weight LLM that classifies for a limited number of activities (basically, just violence; sorry, no scanning for jaywalking or not wearing a mask or drugs or unauthorized manufacturing or illegal immigrants, really just violence) and only records and forwards if that's what it sees. The device has no storage, no ability to load new algorithms, this is all it does. Also, a hash of every recording is published onchain (and a device can't record the next one until it sees confirmation of the previous one) so we can all count how much recording is done. Finally, everyone has a legal right to take any such device they see and grab it and dissassemble and verify the hardware themselves. Explicitly fund university clubs, hacker groups, etc to regularly do this.

First @tedlieu. Now @RepGraceMeng. Both Taiwanese American Democrats. The racial cover isn't even subtle. Rep. Grace Meng calls the @nvidia chip sales a "five alarm fire" and demands full transparency. Fine. Let's start with her. Let voters see how the Washington machine really works. The checks: Three @AnthropicAI employees gave her $17,000 this cycle. The PAC of Americans for Responsible Innovation (@americans4ri) added $2,000. The same donors: Those same three Anthropic employees also funded ARI's PAC—$85,000 in total. One gave Meng $3,000, then nearly $1M to Public First, an AI super PAC network. The same operative: ARI and Public First share a founder: former Democratic congressman Brad Carson (@bradrcarson). Anthropic gave Public First Action $20M, and Dario Amodei (@DarioAmodei) added $1M of his own. The same agenda: Two days after the midterms, ARI led 52 groups urging Democrats to hand out committee chairs based on AI safety positions. The seat and power: With Democrats taking the House, ARI is pushing hard for Grace Meng to chair the Commerce, Justice, Science (CJS) Appropriations subcommittee— the panel that funds the Department of Commerce (@CommerceGov). Commerce runs AI chip export controls and @NIST's AI safety work: Anthropic's top two priorities in Washington. Now she's going after NVIDIA—Anthropic's fiercest rival on export controls—and Commerce, the very agency she'll fund. Same donors. Same operative. Same agenda. Same subcommittee. No law needs to be broken. Interests find power. Power finds interests. That's the five-alarm fire. The real one. Congresswoman, you asked for full transparency. Here it is.

🚨 BREAKING: Usman Nurmagomedov and 5 other members of his team have been sentenced to 6 months in prison, fined $300K AED, and ordered to be deported after serving their sentences Merab and Saba were each fined $50K AED for "disturbing public security and assaulting others" The two people who filmed the brawl were also sentenced to 6 months in prison, fined $300K AED and ordered to be deported for "for violating privacy by using information technology to film and publish footage of the brawl" (via @ADJD_Official)

Elon Musk offered $97.4 billion to buy OpenAI, and Sam Altman answered in under 60 minutes. “no thank you but we will buy twitter for $9.74 billion if you want.” Same digits, 1/10th the price the insult was in the math. Rewind to 2015, when Musk co-founded OpenAI as a nonprofit, put in $44 million and sent late-night emails about saving humanity. Then he walked. OpenAI took $13 billion from Microsoft, shipped ChatGPT and grew into a $135 billion company without him. Altman wrote him a private email: “I don’t think OpenAI would have happened without you and it really fing hurts when you publicly attack us.” Musk reply came in public instead “Scam Altman,” “Greg Stockman,” and then a lawsuit. In court, OpenAI lawyer handed him a 2018 email from his own inbox, spelling out the plan to raise $10 billion years before he claims he learned of it. Musk, under oath: “I did not read the fine print.” The lawyer: “It’s a 4-page document.” The jury ruled against him unanimously. So count what’s left on the table: a $44 million donation, a $97.4 billion bid dead in an hour, and a $135 billion company run by the man he once mentored. He co-founded the future, then sued it and it won.

🚨 BREAKING: Usman Nurmagomedov has been sentenced to 6 MONTHS in prison in Abu Dhabi for attacking Merab Dvalishvili, along with all 5 other members of Team Nurmagomedov They’ve all been fined Dh300K each and will be DEPORTED after serving their sentences via @TheNationalNews

uncensored z-image turbo is 7.2gb and huggingface let anyone download it 😳 its free, apache 2.0 license, no gate, 110k downloads - nsfw finetune of z-image turbo, made for comfyui - creator says it's for adults only and uses the official turbo text encoder and vae - one 7.2gb q8 file is all u need try it here:

I am actually disgusted by the timing of this drop. And this will get a bit more technical. Let’s do a real technical teardown of this "1,385-sigma Blackwell weight fingerprint" report that OpenAI and Anthropic are breathlessly promoting today, because banking on nobody reading past the headline is how DC narratives get manufactured. ​First, look at the actual CUDA maths: when PyTorch initialises model weights (torch.nn.init) with a fixed PRNG seed, curand/Philox parallelises random number generation across the GPU’s Streaming Multiprocessors (SMs). Because Hopper has 132 SMs and Blackwell has 148 or 152 SMs, the exact same seed lays out initial weights in a different spatial pattern depending on the GPU’s SM count. In huge MoE models, certain low-gradient null-spaces or unrouted slices preserve faint directional traces of those step-0 weights even after training. ​So yes, brute-forcing the seed and SM grid can show that a model’s step_0 weight initialisation script ran on a 148/152-SM Blackwell GPU. Now read what the summary explicitly admits in the fine print: 1. Initialisation is not full training: The report literally states that these signatures show how the weights were created, and "support (but do not prove) the inference that Blackwell also participated in model training". Initialising a model checkpoint takes seconds on a single rented dev node before loading the safetensors onto a standard Hopper cluster. The "2,000 to 3,000 Blackwell GPUs" and "$500 million" cluster numbers do not come from the weight fingerprint and Depue admits in his own thread that he simply "estimated" cluster sizes from release schedules and model details. ​2. Zero geolocation and zero proof of smuggling: The final sentence of the executive summary explicitly admits that the findings "do not establish the physical location of the chips, how they were acquired, or whether export control regulations were violated". Depue himself rented a cloud GB200 instance just to run his verification test. Commercial labs that are not on the Entity List renting bare-metal Blackwell nodes in offshore clouds (Singapore, Japan, or the Middle East) is completely legal under BIS rules and leaves the exact same SM fingerprint. Now ask yourself why a former OpenAI researcher dropped this on November 12, the exact day Jensen Huang was pushed out, with OpenAI and Anthropic immediately stepping in to "independently verify" it. US closed-source labs are watching their paid API moats get eaten alive by free Chinese open-weights models. Conflating a step-0 PRNG initialisation seed with a "$500M smuggled Blackwell cluster" hands DC hawks the exact pretext OpenAI and Anthropic want to regulate global cloud compute and kneecap open-source AI.