Keep on to blur preview images; turn off to show them clearly
RT @Jimmy_JingLv: 如果让我颁奖的话，我愿意把 2025 年最好的 AIGC 视频颁给这个！ 那就是以音乐综艺的形式整活，😂太会了啊！ 一切模式，包括前采、演唱、后采、跟观众的互动等等，都是和《我是歌手》一样的模式。要素太多了！… https://t.co/7…
🚧 building https://t.co/AJfZ3LMlgq https://t.co/606cFUoda3 https://t.co/s0m0tpQMDH https://t.co/UQ5vrrYdAG 🐣learning/earning while helping others ❤️making software, storytelling videos 🔙alibaba @thoughtworks
One more issue that Westerners miss is that China isn't just able to pay top talent top buck. It offers them top challenges. For all the significance… ASML/TRUMPF/Cymer/Zeiss are a small pond. Risk-averse, commercial shops. Lin Nan is a scientist. In China, he can play around.
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
RT @andimarafioti: The Reachy mini desktop app is so aesthetically pleasing. Huge fan.
Co-founder & CEO @HuggingFace 🤗, the open and collaborative platform for AI builders
With all of the quantum computers that are supposedly out there already, why hasn't anybody yet repeated the Kwiat et al. 1999 quantum Zeno experiment with a qbit instead of an opaque object?
Why not a Conditional Phase Gate (or Dispersive Readout)? By switching from an opaque object (which blocks/absorbs) to a birefringent object (which shifts phase), we could move from a "survival test" (Zeno) to an "interference test," which is much more robust against decoherence. 1. The Pivot: From "Wall" to "Lens" In the original Kwiat/Zeno experiment, the "Bomb" (Observer $P$) acts as a Wall: The Mechanism: It either stops the photon (absorption) or lets it pass. The Problem: To prove the object is in a superposition, you need the photon to survive a "danger zone" many times ($N \to \infty$). Any imperfection leads to loss, which mimics collapse. Why not replace the "Bomb" with a Birefringent Lens? The Mechanism: The object allows the photon to pass in both states, but it "tags" the photon with a specific phase shift or polarization rotation depending on the object's state. If Object is $|\uparrow\rangle$: Photon passes with Phase $+\phi$. If Object is $|\downarrow\rangle$: Photon passes with Phase $-\phi$. The Result: The photon is never blocked. It always passes through, and the system becomes entangled rather than frozen. 2. Overcoming the Engineering Challenges A. Solving "Low N" (The Single-Shot Solution) The Zeno effect requires $N \approx 20-100$ cycles to "freeze" the state. In a birefringent setup, you only need $N=1$ (a single pass). You do not need to freeze the object's evolution; you only need to correlate with it. As long as the photon interacts once, it carries the phase information. This reduces the experiment time from microseconds to nanoseconds, beating the "decoherence clock." B. Solving "Low Interaction" (Weak Measurements) In the Zeno experiment, if the "bomb" is only 50% opaque, the experiment fails. In this experiment, if the interaction is weak (e.g., the phase shift is tiny, say $10^\circ$), it can still work. Interferometry: You mix the output photon with a reference beam. Even a tiny phase shift creates a detectable change in the interference pattern. Weak Values: By repeating the experiment many times, you can statistically distinguish the interference pattern of a superposition from the pattern of a classical mixture, even with weak interactions. C. Solving "Low Eta" (The Heralding Strategy) This is the most critical part. Can we recover information "with high confidence... despite the detector having clicked." The Solution: We use Heralding (Post-Selection). We place detectors at the end of the path and only analyze the data where the detectors actually click. The Result: If the surviving photons show interference fringes correlated with the qubit, we have proven that—for those specific events—the superposition remained intact. The Caveat: This may satisfy many. But it would not "fully vindicate" RQM against a die-hard. A skeptic could still argue the detection loophole: "The collapse DID happen! It happened in the 50% of photons that were lost (Low Eta). You are only showing me the survivors." To close this loophole completely, we would still need high efficiency ($\eta > 82\%$), but the birefringent setup is much more likely to achieve this than the Zeno setup because passing light through a transparent crystal is inherently less lossy than bouncing it off mirrors. 3. The Specific Experiment Configuration To realize the concept, we would not use "polarizers" in the classical sense. We would use Circuit QED (Superconducting Qubits), which is the leading candidate for this type of "Mesoscopic Superposition." The Setup: Observer ($P$): A Transmon Qubit (an artificial atom visible to the naked eye) inside a microwave cavity. We prepare it in a superposition: $\frac{1}{\sqrt{2}}(|g\rangle + |e\rangle)$. System ($S$): A microwave photon pulse. Birefringence (Dispersive Shift): We tune the cavity to the Dispersive Regime. If the Qubit is $|g\rangle$, the cavity effectively has "Refractive Index A." If the Qubit is $|e\rangle$, the cavity has "Refractive Index B." Detectors: Crucial Correction: You mentioned placing detectors "at the position of the object." This would destroy the superposition. Instead, we place detectors after the cavity. We measure the phase of the reflected photon (Homodyne Detection) and then measuring the state of the Qubit. The Vindication: We look for Bell Correlations between the Photon's phase and the Qubit's state. If we see these correlations, we have proven that the Qubit did not "collapse" to a classical state when the photon hit it; instead, it entered an entangled superposition with the photon. This experiment is technically feasible today and is the strongest candidate for the "sophisticated experiment" Rovelli imagined.
RT @WentaoGuo7: 🚀SonicMoE🚀: a blazingly-fast MoE implementation optimized for NVIDIA Hopper GPUs. SonicMoE reduces activation memory by 45%…
🇦🇺 Co-founder: @AnswerDotAI & @FastDotAI ; Prev: professor @ UQ; Stanford fellow; @kaggle president; @fastmail/@enlitic/etc founder https://t.co/16UBFTX7mo
A very Hajnali moment. Individual greatness > object-level concept. Of course, Soviets too were calling their aerospace design bureaus names like «Sukhoi», «Tupolev» and «Mikoyan».
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1