只有河南跨年结束后的停车场。#只有河南 >>阅读更多

    To their confusion, and often to their regret, Americans seemed faced with two Jimmy Carters. One was the man whose presidency, won by a squeak and relinquished amid the humiliation of the Tehran hostage crisis, seemed an essay in weakness and naivety. This was the chief executive who once addressed the nation in a grey cardigan, sitting by a guttering fire; who, at the peak of the energy crisis in 1979, as Americans queued miserably for petrol, wanly criticised their malaise; who, at peak of the cold war, seemed to hope he could effect a thaw by writing a personal letter to the exiled nuclear physicist Andrei Sakharov; and whose bid to rescue the hostages ended with a helicopter crash in the desert. This was also the man who, out fishing, was said to have been assailed by a “killer rabbit” that swam towards him; who, when jogging, suffered heatstroke; and who admitted to Playboy magazine that he had often committed “adultery in my heart”, inciting a wave of mockery from sophisticates on both coasts.#TheEconomist >>阅读更多

    For years, Jenny had been trying to talk with her mother about something that had been put through the machine repeatedly: the sexual abuse of Alice’s youngest daughter, Andrea, by Gerry, and Alice’s refusal to see the harm that it had done. “She loves and protects the most destructive person of my life,” Andrea had written years earlier. Alice used to shut down when Jenny brought up the subject, but after she got Alzheimer’s, Jenny said, “she didn’t feel invested in that person, Gerry, at all, or in the person she’d been with him. She started to lose that great terror over the truth.” Jenny and her mother had lucid conversations about Andrea’s abuse, which Jenny sometimes recorded, but Alice would forget what had happened a few minutes later. In one conversation, in 2019, Alice exhaled loudly and said under her breath, “How awful.” She looked up at Jenny and said, “It was beastly of me not to get rid of him.” “Did you sort of blame yourself and hate yourself and think Andrea would never love you again, too?” Jenny asked, hoping for more self-reflection. “No, I don’t think it was that,” Alice said. “I don’t know why I didn’t.” She sat in a cushioned chair, wearing a zip-up sweater, a fleece blanket spread over her lap. Then she said in a louder voice, as if finally discovering something solid, “Well, he told me he’d kill himself, of course.” Gerry had said that he couldn’t live without her. “He was in a desperate situation.” “And it’s an empty threat, isn’t it?” Jenny said. “What if Andrea had killed herself?” “Yes, exactly,” Alice said, nodding. “A lot of victims of child abuse do,” Jenny said. Alice held her hand to her forehead. She seemed to be losing track of the emotional center of the conversation. “Does she think about it still?” she asked. “This?” Jenny said. “It’s not something you get over.” “Oh, God. Oh, God,” Alice said, in a high, pained voice, bowing her head and holding it in her hand. #TheNewYorker >>阅读更多

    It was, perhaps, only a matter of time before hostilities broke out in Trump World between the MAGA Internet Warriors and the Elon Musk Tech Bro Caucus. The cause? Immigration. There is populist outrage over immigration. Then there is the nuanced outrage offered by Musk and his ilk over illegal immigration. This week's furor around venture capitalist/Musk friend Sriram Krishnan's appointment to the Trump administration highlighted how the two forms of outrage don't always overlap. The civil war between pixel patriots took place, fittingly, on Musk's social-media platform X between some of the loudest players in Donald Trump's orbit, exposing that the echoverse that gave the world's richest man so much influence with the president-elect isn't in lockstep with him on everything.#WSJ #Elon musk #H1B >>阅读更多

    Value of $1,000 invested a decade ago: Nvidia: $272,235 AMD: $47,190 Tesla: $29,890 Broadcom: $24,390 Microstrategy: $20,820 Netflix: $19,020 Amazon: $14,685 Microsoft: $9,150 Apple: $9,090 Meta: $7,470 Alphabet: $7,225 Adobe: $6,030 Salesforce: $5,620#H1B >>阅读更多

    DeepSeek (Chinese AI co) making it look easy today with an open weights release of a frontier-grade LLM trained on a joke of a budget (2048 GPUs for 2 months, $6M). For reference, this level of capability is supposed to require clusters of closer to 16K GPUs, the ones being brought up today are more around 100K GPUs. E.g. Llama 3 405B used 30.8M GPU-hours, while DeepSeek-V3 looks to be a stronger model at only 2.8M GPU-hours (~11X less compute). If the model also passes vibe checks (e.g. LLM arena rankings are ongoing, my few quick tests went well so far) it will be a highly impressive display of research and engineering under resource constraints. Does this mean you don't need large GPU clusters for frontier LLMs? No but you have to ensure that you're not wasteful with what you have, and this looks like a nice demonstration that there's still a lot to get through with both data and algorithms. Very nice & detailed tech report too, reading through. >>阅读更多