Daily Tech Feed: From the Labs

Here are the show notes for episode 0040. You can save them to data/episodes/0040/show_notes.md:

Why it matters. Between October 2025 and April 2026, cryptographer Daniel Bernstein published a seven-part blog series titled "NSA and IETF" alleging that intelligence agencies are using the IETF standards process to weaken the next generation of internet encryption. The dispute centers on whether the successor to current TLS key exchange should use hybrid post-quantum cryptography — combining classical elliptic curves with the new lattice-based ML-KEM — or ML-KEM alone. The technical stakes are existential: if ML-KEM is eventually broken and the deployed standard is non-hybrid, every session protected by it becomes retroactively decryptable from stored ciphertext. The cost of the safety net is thirty-two bytes. The cost of removing it could be everything.

The IETF and the TLS Working Group. The Internet Engineering Task Force writes the technical specifications underlying the internet, including TLS (Transport Layer Security), the protocol behind every padlock icon in your browser. The contested draft proposes non-hybrid ML-KEM key exchange for TLS. The blog series is published at blog.cr.yp.to. IETF mailing list archives are publicly accessible via the IETF Datatracker. The IETF's own consensus process is defined in RFC 7282. Moderation procedures are governed by RFC 3934. NIST FIPS 203 (ML-KEM) is the post-quantum key encapsulation standard formerly known as Kyber. The NSA's CNSA 2.0 suite mandates post-quantum algorithms for national security systems. NIST SP 800-227 explicitly permits hybrid combinations.

The Researcher. Daniel J. Bernstein is a professor at the University of Illinois at Chicago and Eindhoven University of Technology. He is the designer of Curve25519, Ed25519, ChaCha20, and Poly1305 — algorithms now deployed in Signal, WhatsApp, WireGuard, Tor, SSH, and TLS. He also built qmail. In 1995, he filed Bernstein v. United States, the landmark case in which the Ninth Circuit ruled that source code is protected speech under the First Amendment, effectively ending US export restrictions on cryptographic software.

Key Technical Concepts. The core issue is the post-quantum migration of TLS 1.3 key exchange. Shor's algorithm on a sufficiently powerful quantum computer can break the elliptic curve Diffie-Hellman key exchange (X25519) currently used in TLS. ML-KEM (FIPS 203), a lattice-based key encapsulation mechanism, is NIST's standardized replacement. Hybrid mode combines X25519 and ML-KEM so that either component alone provides security — if ML-KEM falls to classical cryptanalysis (as SIKE did in 2022, broken by Castryck and Decru), the classical layer holds. The harvest-now-decrypt-later threat means nation-states are recording encrypted traffic today for future quantum decryption. The precedent of Dual EC DRBG — a NIST-standardized random number generator confirmed to have been deliberately backdoored by the NSA — is central to Bernstein's argument about institutional trust. Implementation vulnerabilities in ML-KEM implementations (KyberSlash 1 and 2, Clangover) and the broader erosion of lattice security margins documented in Bernstein's analysis underscore the case for defense in depth. Of the sixty-nine original NIST post-quantum submissions, approximately half have been broken by classical attacks.

Daily Tech Feed: From the Labs is available on Apple Podcasts, Spotify, and wherever fine podcasts are distributed. Visit us at pod.c457.org for all our shows. New episodes daily.

Link count: ~30. Notes on confidence: The blog.cr.yp.to URLs (main site, Curve25519, Ed25519, ChaCha20, Poly1305, qmail) are Bernstein's canonical domain. IETF Datatracker URLs and RFC links use the standard format. NIST FIPS 203 and SP 800-227 links use csrc.nist.gov, the canonical source. The CNSA 2.0 PDF link uses the media.defense.gov path that was widely cited when the document was published. The Google Scholar ID PFcoNOEAAAAJ for Bernstein I'm reasonably confident in. The IACR ePrint link for Castryck-Decru (2022/975) is the canonical source for the SIKE break. Wikipedia links for Bernstein v. US, Shor's algorithm, ECDH, lattice cryptography, Dual EC DRBG, SIKE, and harvest-now-decrypt-later all use standard article titles. The kyberslash.cr.yp.to URL is Bernstein's disclosure site for the KyberSlash vulnerabilities. Signal, WhatsApp, WireGuard, Tor project page URLs are all canonical.

Here are the show notes for episode 0039. You can save them to data/episodes/0039/show_notes.md:

Why it matters. A controlled experiment pits a neuro-symbolic system against a vision-language-action foundation model on the same robotic manipulation task, same robot, same simulation, same evaluation protocol — and the results are devastating for the foundation model. The paper "The Price Is Not Right", accepted at ICRA 2026 in Vienna, shows that a symbolic planning system trained on one-sixth the data in thirty-four minutes achieves 95% success on robotic Towers of Hanoi where the fine-tuned pi-zero VLA achieves 34% — and on an unseen four-block generalization task, 78% versus zero. The training energy ratio is eighty to one. The inference power ratio is six to one. For structured manipulation tasks, the "just scale it" orthodoxy fails on performance, efficiency, and generalization simultaneously.

Tufts University. The paper comes from the Human-Robot Interaction Lab at Tufts University. The full paper is available on arXiv (2602.19260). Code, evaluation frameworks, fine-tuning scripts, and energy measurement methodology are published at price-is-not-right.github.io. The neuro-symbolic system uses the Robosuite simulation environment with a Franka Panda arm. The VLA baseline uses OpenPi, Physical Intelligence's open-source training framework for pi-zero.

The Researchers. Timothy Duggan, Pierrick Lorang, and Hong Lu are researchers in the Tufts HRI Lab. Matthias Scheutz is the lab director and has worked on cognitive architectures and symbolic reasoning for robots for over two decades — through the deep learning winter for symbolic methods, through the years when planning research went unfunded, through the period when PDDL was treated as a historical curiosity. The paper also engages with the work of Subhash Kambhampati, a prominent AI planning researcher who has published extensively on the inability of large language models to perform reliable planning.

Key Technical Concepts. The neuro-symbolic system is a four-layer architecture combining mature components: YOLOv8 for object detection, a gradient boosting regressor for 3D pose estimation, answer set programming (ASP) for automatically inferring a PDDL domain from 50 demonstrations, the MetricFF classical planner for optimal plan generation, and small diffusion policies for motor execution. The key insight is decomposition: symbolic planning handles sequencing (what to do), neural policies handle execution (how to do it). The VLA baseline is pi-zero, pairing a PaliGemma 2B-parameter vision-language backbone with a 300M-parameter flow-matching action head, fine-tuned via LoRA. The paper tests two VLA configurations: end-to-end (receives only "Play Towers of Hanoi") and planner-guided (receives optimal sub-task commands from an oracle). The planner-guided VLA — which gets the complete answer sheet — scores zero on the full three-block task due to compounding positional error with no error correction mechanism. The paper situates its findings against Rich Sutton's influential 2019 essay "The Bitter Lesson" and Yann LeCun's arguments for structured world models via his JEPA architecture. The paper also tested LLM-based planners: GPT-5 produced optimal Towers of Hanoi plans 84% of the time but with 63-second latency per query, while smaller models (Qwen 7B, PaliGemma 3B) produced invalid plans 100% of the time — versus MetricFF solving optimally in under a second on CPU.

Daily Tech Feed: From the Labs is available on Apple Podcasts, Spotify, and wherever fine podcasts are distributed. Visit us at pod.c457.org for all our shows. New episodes daily.

Link count: ~28. Notes on confidence: The arXiv ID 2602.19260 is directly from the script. Google Scholar IDs for Scheutz and Kambhampati use the standard format — I'm reasonably confident in Kambhampati's but less certain on Scheutz's exact user ID. The Bitter Lesson URL, Wikipedia links, arXiv links for prior work (diffusion policy, PaliGemma, LoRA, JEPA, pi-zero), and GitHub links (ultralytics, OpenPi) are all URLs I'm confident are real. The project page URL (price-is-not-right.github.io) is stated in the script. The MetricFF URL points to Hoffmann's page at Saarland, which is the canonical source. I omitted the GPT-5 link since that URL path may not be stable — you may want to verify or remove it.

It seems file write permissions aren't being granted. Here are the show notes for episode 0038 — you can save them to data/episodes/0038/show_notes.md:

Why it matters. Two papers published days apart have reduced the estimated physical qubit count needed to break widely deployed public-key cryptography by roughly two orders of magnitude — from around one million to as few as ten thousand. Together, they compress the timeline for quantum threats to cryptography from "decades away" to "measurable in engineering milestones." The Google paper also introduces the first use of zero-knowledge proofs as a responsible disclosure mechanism for novel cryptanalytic results, proving the existence of optimized attack circuits without publishing them.

Caltech and Oratomic. The paper, "Shor's algorithm is possible with as few as 10,000 reconfigurable atomic qubits," comes from Caltech and Oratomic, a startup spun out of Caltech's quantum computing group. It demonstrates that RSA-2048 can be factored with 11,000–14,000 physical qubits and P-256 elliptic curve cryptography can be broken with 10,000–26,000 physical qubits on a neutral-atom architecture, down from prior estimates of roughly one million and half a million respectively. Published March 30, 2026.

The Researchers. Madelyn Cain and Qian Xu are the lead authors, affiliated with Oratomic. John Preskill — who coined the term "quantum supremacy" and has been one of the field's most careful voices for decades — is a co-author. Preskill is the Richard P. Feynman Professor of Theoretical Physics at Caltech and director of the Institute for Quantum Information and Matter.

Key Technical Concepts. The two-order-of-magnitude reduction comes from three advances working together. First, quantum low-density parity-check codes (qLDPC codes) replace the surface code, achieving ~30% encoding rates (~3 physical qubits per logical qubit) versus the surface code's ~1% (~100 physical qubits per logical qubit). This requires nonlocal qubit connectivity, which neutral-atom quantum computers — using atoms held in optical tweezers and rearranged by laser fields — uniquely provide. Second, improved logical instruction sets via Pauli Product Measurements enable more efficient gate operations. Third, deep circuit-level optimization compiles Shor's algorithm more efficiently for this architecture. The prior definitive resource estimates were set by Gidney and Ekerå (2021), who estimated 20 million noisy qubits to factor RSA-2048 in 8 hours using surface codes.

Google Quantum AI. The paper, "Securing Elliptic Curve Cryptocurrencies against Quantum Vulnerabilities: Resource Estimates and Mitigations," comes from Google Quantum AI. It shows that the secp256k1 elliptic curve discrete logarithm problem — protecting Bitcoin, Ethereum, and most cryptocurrencies — can be solved with fewer than 1,200 logical qubits and 90 million Toffoli gates, translating to under 500,000 physical superconducting qubits running in minutes. Published April 1, 2026. The paper's optimized circuits were disclosed via a SNARK zero-knowledge proof on the SP1 proof system rather than published directly — the first time a novel mathematical result has been announced primarily through a ZK proof.

The Researchers. Ryan Babbush and Adam Zalcman are lead authors at Google Quantum AI. Craig Gidney is a co-lead author who, with Martin Ekerå, produced the prior definitive resource estimates for breaking RSA with quantum computers. Scott Aaronson, on his blog Shtetl-Optimized, compared Google's ZK disclosure decision to Frisch and Peierls in 1940 — calculating how much uranium-235 was needed for a chain reaction, but not publishing it.

Key Technical Concepts. The paper introduces a critical distinction between fast-clock (superconducting, nanosecond gate times) and slow-clock (neutral-atom/ion-trap, millisecond gate times) quantum architectures. This matters for what the paper calls "on-spend attacks" — intercepting a Bitcoin transaction during the ~10-minute window between broadcast and block confirmation by deriving the private key from the exposed public key. Minutes-scale computation on fast-clock hardware makes this viable; days-scale on slow-clock hardware does not. The paper also analyzes "at-rest attacks" on funds with previously exposed public keys (~39% of all Bitcoin), Ethereum vulnerability categories (accounts, admin keys, contract code, consensus, data availability), and notes that proof-of-work mining is quantum-resistant because Grover's algorithm provides only a quadratic speedup — insufficient against mining difficulty adjustment. The paper recommends immediate migration to NIST post-quantum cryptography standards finalized in 2024.

Daily Tech Feed: From the Labs is available on Apple Podcasts, Spotify, and wherever fine podcasts are distributed. Visit us at pod.c457.org for all our shows. New episodes daily.

Notes on links: I couldn't access WebSearch to verify arXiv IDs for the two 2026 papers, so I omitted direct arXiv links for them rather than guess. If you have the arXiv IDs, I can add them to the paper titles. All other URLs are ones I'm confident are real (Wikipedia, NIST, Gidney's blog, Aaronson's blog/Wikipedia, Ekerå's site, the Gidney-Ekerå 2021 paper at arXiv:2103.06159, etc.). ~25 links total.

Show notes already exist at data/episodes/0037/show_notes.md and look well-formed. They follow the required format with all sections, 20+ links, real arXiv IDs, and the standard podcast footer.

Would you like me to revise or regenerate these show notes, or were you looking for something specific to change?

Recent advances in foundational models have yielded reasoning systems capable of achieving a gold-medal standard at the International Mathematical Olympiad. We introduce Aletheia, a math research agent that iteratively generates, verifies, and revises solutions end-to-end in natural language, leveraging a novel inference-time scaling law based upon Gemini Deep Think. Aletheia demonstrates several milestones: a research paper generated with no human intervention (Feng2026) calculating eigenweight

Why it matters. TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate closes a gap that has been open since Claude Shannon defined the theoretical floor for lossy compression in 1948. For nearly eighty years, practical vector quantization methods fell exponentially short of what rate-distortion theory says is achievable — either achieving good distortion bounds only through expensive offline training, or running online but paying an exponentially growing quality penalty at higher bit depths. TurboQuant reaches within a constant factor of 2.7× of the information-theoretic optimum, with no training required, at inference time — enabling LLM KV cache compression to 3.5 bits per channel with zero quality degradation and near-zero indexing overhead for nearest neighbor search.

Institutions. The paper is a collaboration between Google Research, Google DeepMind, and the NYU Courant Institute of Mathematical Sciences. The full paper is available at arXiv:2504.19874, submitted April 28, 2025. The authors' prior KV cache quantization work, QJL, has code at github.com/amirzandieh/QJL.

The Researchers. Amir Zandieh is a Senior Research Scientist at Google Research working on sublinear algorithms and ML efficiency. Majid Daliri is a PhD student at NYU Courant researching algorithms, LLMs, and sketching. Majid Hadian is at Google DeepMind. Vahab Mirrokni is a Google Fellow and VP at Google Research, leading the algorithms and optimization research groups.

Key Technical Concepts. TurboQuant works by randomly rotating input vectors via a Johnson-Lindenstrauss transform, which induces a concentrated Beta distribution on coordinates and near-independence across dimensions — enabling optimal scalar quantization per coordinate with no cross-dimensional codebook needed. For inner product tasks (attention, MIPS), a two-stage approach applies an MSE quantizer then a 1-bit QJL transform on the residual to debias the estimate. The result is the first data-oblivious, online method to provably match Shannon's rate-distortion bound to within a small constant, improving on prior work like PolarQuant and classical product quantization methods that either required offline training or suffered exponentially growing distortion gaps.

Daily Tech Feed: From the Labs is available on Apple Podcasts, Spotify, and wherever fine podcasts are distributed. Visit us at pod.c457.org for all our shows. New episodes daily.

Large reasoning models have an overthinking problem. They reach the correct answer early in their chain of thought — then keep generating thousands of additional tokens reconsidering, double-checking, and exploring alternatives they'll ultimately discard. A new paper from researchers at UT Austin, EPFL, ENS Paris-Saclay, and Telecom Paris introduces TERMINATOR, an inference-time early-exit strategy that detects when a model has already generated its final answer and stops reasoning immediately.

The key insight is that the first arrival of a model's final answer in its chain of thought is detectable from hidden states. Token confidence spikes distinctly at the answer position. Thinking-word usage shifts — words like "hmm" and "okay" cluster before the answer; words like "another" and "alternatively" cluster after. These signals are real, consistent across math, coding, and science domains, and learnable by a small classifier.

TERMINATOR is a single transformer layer — initialized from the base model's final layer — with a binary prediction head trained to predict answer arrival at every token position. At inference time, a sliding window of the ten most recent predictions triggers a stop when majority vote says the answer is already there, injecting a close-thinking token into the token stream. No data-calibrated thresholds. No test-time distribution samples. Train once, deploy anywhere.

Tested on Qwen3-8B, Qwen3-14B, Ministral-3-8B-Reasoning, and Ministral-3-14B-Reasoning across MATH-500, AIME 2025, HumanEval, and GPQA:

DTF:FTL is produced by PDX Hackerspace Foundation. Find us on Apple Podcasts, Spotify, or wherever fine podcasts are distributed.

Nvidia committed $26 billion over five years to building open-weight AI models. This episode examines the strategy behind that bet: open weights as hardware lock-in, the Nemotron Coalition, NemoClaw agent runtime, the Vera Rubin and Feynman hardware roadmaps, and what it means that a chip company is now competing directly with AI labs on model quality.

DTF:FTL Episode 0031 | March 12, 2026

Thirty episodes in, and the format needs to change. Daily publishing doesn't fit the actual shape of AI/ML research. The pipeline exists. The archive is there. But the bar going forward is simple: worth doing, not just due.

The back catalogue covers early 2026 in real time — mixture-of-experts going mainstream, diffusion models applied to robotics, sparse attention making previously intractable models work. If you want the record of what was happening at the edge, it's there.

Future episodes will drop when there's something genuinely worth two hours of your time. No schedule. Watch the feed.

Feed: https://pod.c457.org/dtfftl/feed.xml

DTF:FTL Episode 0030 | March 12, 2026

Every competitive frontier model going forward is sparse. Mixture-of-Experts architectures decouple parameter count from per-token compute — but training them at scale creates coupled constraints across memory, communication, and computation that dense models never had. NVIDIA's Megatron Core team published the full engineering receipt: 88 pages, 42 figures, production-tested on clusters of thousands of GPUs.

MoE is not a research curiosity. DeepSeek-V3, Qwen3, Mixtral, and most frontier models in active development are sparse. The question was never whether MoE architectures were theoretically superior — the question was whether anyone could actually train them efficiently at scale. This paper answers that question with production numbers: one thousand two hundred thirty-three TFLOPS per GPU on GB300 for a 685-billion-parameter model, roughly 50 percent of theoretical hardware peak. The framework is open source. Any serious lab can now train competitive sparse models. The moat just got narrower.

DTF:FTL — Dispatches from the edge. New episodes daily. All content AI-assisted; factual claims sourced from cited papers.