AI isn’t just changing what startups build, it’s changing how the companies themselves operate. Floodgate co-founder and Partner Ann Miura-Ko explains her thesis around the “AI-pilled organization,” where agents are deployed across engineering, sales, marketing, and strategy to dramatically accelerate how companies learn and execute. She joins Ed Ludlow on "Bloomberg Tech." -------- Like this video? Subscribe to Bloomberg Technology on YouTube: https://www.youtube.com/channel/UCrM7B7SL_g1edFOnmj-SDKg
Bloomberg’s Ed Ludlow explains why Meta agreed to pay up to $18 billion in landmark settlements to resolve allegations that the company deliberately designed Facebook and Instagram to encourage compulsive use among young users. Plus, Nvidia reports after the closing bell today. And, Southern Company CEO Chris Womack joins to discuss the AI boom's impact on the American power grid.
00:00:00 - Bloomberg Tech Begins 00:01:13 - Ian King, Bloomberg News 00:02:31 - Jordan Klein, Mizuho Securities 00:09:38 - JB McCuskey, West Virginia Attorney General 00:27:01 - Chris Womack, Southern Company 00:34:37 - Ann Miura-Ko, Floodgate 00:40:40 - Alexandra Levine, Bloomberg News 00:44:30 - Kunjan Sobhani, Bloomberg Intelligence -------- "Bloomberg Technology" is our daily news program focused exclusively on technology, innovation and the future of business hosted by Ed Ludlow from San Francisco and Caroline Hyde in New York.
SGLang brings Day-0 support for Qwen 3.8-Flash-Next, an early preview of the Qwen4 architecture!
In collaboration with @Alibaba_Qwen, @NVIDIAAI , and @AIatAMD, we've optimized this 125B MoE (6B active, plus 51B N-gram embeddings) for efficient inference:
SGLang brings Day-0 support for Qwen 3.8-Flash-Next, an early preview of the Qwen4 architecture!
In collaboration with @Alibaba_Qwen, @NVIDIAAI , and @AIatAMD, we've optimized this 125B MoE (6B active, plus 51B N-gram embeddings) for efficient inference:
Agentic LLM pipelines face escalating inference costs as context accumulates across retrieval, tool use, and multi-turn interactions. To control latency, deployments routinely compress inputs, but this degrades task accuracy. Speculative decoding (SD) accelerates generation losslessly, yet it assumes the drafter and verifier share an identical context, preventing SD from resolving the accuracy-overhead trade-off. We propose AsymSpec, an asymmetric speculative decoding framework that breaks this symmetry: a lightweight drafter reads the full input while the large verifier operates on the compressed view. The drafter steers the verifier via a contrastive $δ$-fusion of logits, modulated by a divergence-aware acceptance gate that preserves verification stability and high draft acceptance rates. Evaluated across four agentic capabilities and two end-to-end agent benchmarks, AsymSpec reaches $\approx 90\%$ of full-context accuracy on average, delivering $1.3$--$1.7\times$ throughput speedups at $0.2$--$0.3\times$ the compute cost on isolated text capabilities. These results show that asymmetric context access yields substantial gains precisely when compression discards critical reasoning signals.
Orthogonal optimisers such as Muon can substantially accelerate large language model pretraining relative to Adam, yet the mechanism remains incompletely understood. We investigate this through an out-of-sample spectral probing analysis of Transformer loss landscapes. At checkpoints along real training trajectories, we decompose each momentum buffer into its singular directions and estimate the loss-optimal step size along each direction on held-out data. The resulting spectral profile is anisotropic yet stable across batches and training stages, and consistent across the optimisers and model scales: a volatile head operating at the Edge-of-Stability supports a much smaller step size than the tolerant bulk, which permits substantially larger steps. This profile provides a unified spectral allocation account of why Muon outperforms Adam, which outperforms SGD. It also exposes a limitation of Muon's uniform scaling: it still underutilises the bulk. Guided by this finding, we introduce Spectral-Aware Muon (SAMuon), which holds the head at the Muon scale and amplifies the bulk using a static spectral prior. We provide two variants: the complete SAMuon follows the measured profile using a low-rank randomised SVD and the simplified SAMuon-lite uses a two-level approximation via rank-one power iteration. Neither method adds persistent optimiser state or notable extra FLOPs beyond Muon at scale, and the idealised exact-whitening versions of both retain Muon's asymptotic convergence rate under standard assumptions. Across "modded-nanogpt" models from 124M to 1B parameters, both variants outperform tuned AdamW and Muon (Scion implementation) baselines in all evaluated model-scale and batch-size configurations. SAMuon requires 13.3% to 24.0% fewer training tokens to reach the same validation loss as Muon, while SAMuon-lite retains most of this gain with near-zero wall-clock overhead.
For the first time, we’ve given external researchers a way to study AI’s impacts using real, privacy-preserved Claude usage data. To date, this work has only been possible within AI labs. We can’t tell the whole story alone, so we opened up our tools.https://t.co/suLJcARAu1
For the first time, we’ve given external researchers a way to study AI’s impacts using real, privacy-preserved Claude usage data. To date, this work has only been possible within AI labs. We can’t tell the whole story alone, so we opened up our tools.
Hitting Claude's limit mid-work is annoying af. Diet Claude gives a live usage meter shows how much of your session you’ve used, how much time is left, and when your limits reset. It helps optimise token usage by context trimming, tightening prompts, suggesting apt models. And when you do run dry, it carries your conversation and context over to another LLM instead of starting from scratch.
As the AI boom is driving one of the biggest data center buildouts in history, Southern Company CEO Chris Womack discusses the utility’s growing pipeline of data center projects, its 25-year agreement supporting OpenAI’s planned Georgia expansion, and why he says hyperscalers can bear the cost of new infrastructure while putting downward pressure on electricity rates for existing customers. He joins Ed Ludlow on "Bloomberg Tech." -------- Like this video? Subscribe to Bloomberg Technology on YouTube: https://www.youtube.com/channel/UCrM7B7SL_g1edFOnmj-SDKg
Very excited to see native vision come to the @Zai_org GLM family, and impressed by the architecture work behind it!
With SGLang features like RadixAttention and HiCache, GLM-5.3-Flash is one of the most efficient and capable choices for agentic work across coding and… https://t.co/d1pE9tkrpt
Very excited to see native vision come to the @Zai_org GLM family, and impressed by the architecture work behind it!
With SGLang features like RadixAttention and HiCache, GLM-5.3-Flash is one of the most efficient and capable choices for agentic work across coding and
Multimodal instruction-following models require training data that is accurate, diverse, verifiable, and challenging. Existing synthesis pipelines typically follow a one-pass generate-and-filter paradigm, discarding feedback from failed samples, verifier outcomes, and target-model errors. We present VISA (Visual Instruction Synthesis Agent), an agentic framework that reformulates multimodal instruction synthesis as a self-evolving loop. At each round, VISA analyzes an image to filter incompatible constraints and discover new verifiable ones, samples diversity- and difficulty-aware constraint sets from persistent memory, generates candidate instructions, and verifies the resulting samples with executable tools and structured large language model judges. Failed samples trigger diagnostic-guided recovery, while accepted samples are probed against the target model to estimate difficulty. The resulting verifier signals and target-model failure profiles are written back to memory, allowing subsequent rounds to adaptively expand the constraint space, reduce template repetition, and focus on unresolved model weaknesses. The same verifier contracts further provide reward signals for reinforcement learning without a separately trained reward model. Experiments on MM-IFEval show that VISA consistently improves multimodal instruction following over strong baselines, while preserving general multimodal capability across seven public benchmarks.
Reasoning in language allows foundation models to spend more test-time compute on hard problems, such as those requiring decomposition, constraint tracking, and prediction of future consequences. Whether this mechanism can improve robotic manipulation remains unclear, where long-horizon tasks require tracking partial progress, reasoning about object relations, recovering from mistakes, and steering noisy low-level policies. In this paper, we study whether VLMs can be trained to reason directly in natural language to guide low-level manipulation policies. We introduce $R^3$, a simple post-training recipe that turns off-the-shelf VLMs into robotic reasoners: it first mid-trains a VLM on expert-generated reasoning traces to initialize the desired reasoning style, then improves the reasoner with single-step rubric-based RL from offline action data. Unlike prior robotic reasoning methods that mostly use structured traces as auxiliary supervision, $R^3$ trains free-form language reasoning to produce test-time guidance for action. We instantiate $R^3$ on Language Table and simulated bimanual grocery packing, two controlled testbeds for studying robotic reasoning and long-horizon manipulation. $R^3$ improves exploration and generalization across unseen tasks and significantly outperforms instruction-only imitation learning baselines on both benchmarks. Our analyses suggest that free-form language reasoning can function as a test-time compute mechanism for steering low-level policies. Our project page is available at https://robotic-reasoner.github.io/.
Native visual reasoning treats visual generation as the medium of reasoning itself: visual states (i.e. images and videos) are not merely inputs to be understood or outputs to be rendered, but first-class substrates for problem solving beyond language. Yet progress remains bottlenecked by the lack of scalable training tasks, reliable feedback, and controlled comparisons across generative substrates. In this work, we introduce VBVR-Pro, a closed-loop testbed that makes native visual reasoning through generation trainable, verifiable, optimizable, and experimentally controllable. 1) Task scaling. VBVR-Pro turns visual reasoning into a controlled task space of 300 procedurally generated tasks. Models trained on VBVR-Pro show strong transfer beyond the proposed suite across seven external visual reasoning benchmarks such as RISE-Video, MME-CoF-Pro, and BabyVision. 2) Verifiable rewards. VBVR-Pro provides verifiable reward scorers for task-grounded evaluation. Through a systematic study of leading MLLMs as judges, we identify recurring failure modes of the prevalent VLM-as-a-judge paradigm. In contrast, the proposed scorers are grounded in deterministic, task-specific rules, achieve fine-grained alignment with human judgments. Importantly, they serve as reliable reward signals for large-scale multi-task reinforcement learning and demonstrate stronger post-RL performance across visual reasoning tasks. 3) Mechanism study. VBVR-Pro enables controlled modality studies across more than 30 image, video, and interleaved generators. Our analysis shows that video generation remains strongest for tasks requiring persistent spatiotemporal state tracking, while interleaved generation provides a compute-efficient alternative. Critically, ablations and probing suggest the presence of vision-native trajectories that are crucial to visual reasoning. We release all data, models, scorers, and code.
NVIDIA (NASDAQ: NVDA) today reported revenue for the second quarter ended July 26, 2026, of $96.2 billion, up 18% from the previous quarter and up 106% from a year ago.
We have conducted a thorough investigation into the Hugging Face incident.
We are releasing a technical report and accompanying blog post that reconstruct the agents’ activity, explain why existing safeguards failed, and detail how we’re preventing recurrence.…
We have conducted a thorough investigation into the Hugging Face incident.
We are releasing a technical report and accompanying blog post that reconstruct the agents’ activity, explain why existing safeguards failed, and detail how we’re preventing recurrence.
Agnost AI analyzes conversations between users and your production AI agents and discovers: silent failures, agent behavior drift, hallucinations, user frustration, hidden feature requests, and churn signals. It groups them into recurring patterns, shows the exact users and conversations behind each insight, and turns them into evals and fixes.
Private company data with 4x the depth and 2x the coverage of PitchBook, plus 100+ event signals and news across companies, industries, and topics. Source and diligence deals or make outreach lists and trigger outbound. Built for financial services and GTM teams, pay-as-you-go.
At Hot Chips 2026, Intel detailed more about its Crescent Island AI accelerator, which uses the Xe3P architecture. The accelerator will use liquid-cooled chips and HBM4 memory to serve inference workloads in data centers.
The built in MCP Connectors Catalog expanded to offer 44 new connections for apps you use every day, with easy one click setup in the Hermes Desktop App https://t.co/zSOSMvX4U7
The built in MCP Connectors Catalog expanded to offer 44 new connections for apps you use every day, with easy one click setup in the Hermes Desktop App
To get content containing either thought or leadership enter:
To get content containing both thought and leadership enter:
To get content containing the expression thought leadership enter:
You can enter several keywords and you can refine them whenever you want. Our suggestion engine uses more signals but entering a few keywords here will rapidly give you great content to curate.
Your new post is loading...
![[2608.26004] AsymSpec: Context-Asymmetric Speculative Decoding for Agentic LLMs | AI News | Scoop.it](https://img.scoop.it/R3Nf-aV-zFCPDr3s1jV_-zl72eJkfbmt4t8yenImKBVvK0kTmF0xjctABnaLJIm9)
