AI News

Your Research Assistant. Available wherever you are: 🤖 http://pplx.ai/android // 📱 http://pplx.ai/iphone

Home » Press Releases Tower Semiconductor to Highlight its High-Volume Silicon Photonics and SiGe Solutions for AI Infrastructure, Telecom and Emerging Applications at ECOC 2026 MIGDAL HAEMEK, Israel, September 15, 2026 – Tower Semiconductor (NASDAQ/TASE: TSEM), the leading foundry for high-value analog semiconductor solutions, today announced its participation in upcoming ECOC 2026, taking place September 21–23, […]

SEOUL, September 15 (AJP) -South Koreas humanoid robots hardly look like formidable competitors beside their free-wheeling Chinese counterparts, and their awkward moves betray a much deeper problem.Despite South Koreas much-hyped push to become an artificial intelligence powerhouse, its physical AI ecosystem remains fledgling, short of both scal

New blog post:

We had a million lines of Python to clean up. On September 2nd @Teknium asked Hermes Agent to do it.

1,393 subagents and nineteen hours later, the codebase was 34.4% smaller, saving us nearly $2m in engineering hours.

https://t.co/eH6Gy8RAC5

minimax h3 is now live on together ai

the 33b omni-modal video model generates 4–15s clips at up to 2k with native stereo audio, using text, images, video + audio as context

To answer questions outside of their pre-training data, large language models (LLMs) need access to new information, which can be presented in the context window as documents, encoded into the model's parameters, or injected as latent representations. However, each of these methods comes with different efficiency, cost, and performance trade-offs, with no single winner. We present a controlled comparison of representation-based (KV-cache based) and parametric (fine-tuning-based) adaptation methods on five knowledge-intensive benchmarks. We show that in the oracle setting, Cartridges (KV) are the most accurate injection method at nearly every storage budget, outperforming parametric methods by 10 points. Compaction (KV) matches Cartridges only at low compression rates, lagging behind the parametric methods by 10 points at rates higher than $50\times$. In the more realistic multi-document retrieval scenario, Cartridges are the only method that matches in-context learning (ICL), leading the parametric methods by 29 points and Compaction by 15 points. Nonetheless, Cartridges are also the only method, besides full fine-tuning and large MLP adapters, that suffers from catastrophic forgetting, i.e., a 6% performance degradation on control benchmarks, with 13% in coding.

Large language models can produce fluent answers when their factual support is weak. This paper introduces Chain-of-Self-Questioning (CoSQ), a prompt-only framework that makes answer commitment conditional on an explicit assessment of the information required to answer a question. We evaluate three CoSQ variants under seventeen conditions on the 817-item TruthfulQA multiple-choice validation set using eleven open-weight and hosted model families. In the final balanced-option protocol, Grounded-CoSQ at τ=0.90 reduces the mean unconditional wrong-commitment rate from 13.1% under chain-of-thought prompting to 8.9%, a 32.1% relative reduction, while increasing answered accuracy from 86.9% to 89.7% and answering 87.6% of questions. Both improvements hold for all eleven models and at every evaluated threshold. Critical-CoSQ and Adaptive-CoSQ provide neighboring operating points with 88.6% and 86.5% coverage, respectively, while remaining more reliable than the baseline. A secondary Natural Questions Short-Answer evaluation provides convergent open-form evidence. These findings show that self-assessment can support explicit, tunable answer-or-abstain decisions when an unsupported commitment is more costly than referral or review.

Factory has raised $200 million at a $5 billion valuation, more than tripling the $1.5 billion valuation investors assigned it just five months ago. Blackstone, Khosla Ventures, Sequoia Capital, Insight Partners, Evantic Capital and Sound Ventures backed the new round, with Salesforce CEO Marc Benioff and hedge-fund manager Brad Gerstner participating as angel investors. The

Sep 15, 2026

Startup funding opened the week with a surprisingly coherent message: investors are backing AI where it sits inside hard financial, clinical, security, and data infrastructure rather than financing another wave of generic model wrappers. Hong Kong fintech Qupital announced $300 million in combined new capital commitments, anchored by a Series C and supplemented with asset-backed

Banger paper from Meta.

This work shows that byte-level models start out behind token models and then pass them as compute grows.

They show this for distilled 1B models trained on up to 1 trillion bytes.

To distill a byte student from a token teacher, they convert the teacher's token logits into…

SWE-2 is Cognition's new coding model, post-trained from Kimi K3 with RL that optimizes for cost and capability at the same time. It hits 50.0% on FrontierCode 1.1 Main, within a point of Fable 5.1 at 64% less, and lands within a few points of GPT-6 Astra at a quarter of the cost. Compared to SWE-1.7 it takes 58% fewer turns and costs 81% less while scoring higher. Available now in Devin Desktop and CLI.

Nvidia CEO Jensen Huang pushed back on calls to give AI companies antitrust exemptions to coordinate slowing the development of advanced models.

20GW sits inside a restricted local boundary, 1,525MW actually slips, 2.3GW nationwide including New York

Vidu S2

Real-Time Interactive, Editable, and Spatial Video Generation

paper: https://t.co/K6anO4qxEc

Large language model (LLM) distillation aims to transfer the capabilities of a powerful teacher to a smaller student. Direct imitation, however, can also transfer the teacher's systematic bias and errors. This challenge is particularly pronounced under covariate shift, when the teacher's reliability on target questions is uncertain and target-domain reward feedback is unavailable. We propose Coupled Calibration and Learning (CCL), an LLM distillation algorithm that couples teacher calibration with student updates through token-level branching, using reward feedback only on source questions. Each iteration calibrates the teacher using source feedback and then uses the calibrated teacher to train the student on target questions. The updated student, in turn, informs subsequent calibration. In an autoregressive policy framework, we prove that the output student's expected average Kullback-Leibler divergence to the oracle student converges to zero at a polynomial rate in the number of iterations. The oracle maximizes the true reference-regularized target reward within the student class, which need not represent the unrestricted optimal policy. Our analysis quantifies the progress of projected student gradient updates while controlling the error in teacher calibration. We further establish a separation from regularized direct matching: its error relative to the oracle student can remain bounded away from zero even when the teacher achieves higher regularized target reward than every student policy. These results demonstrate that LLM distillation can overcome persistent teacher bias and recover the optimal student through coupled calibration and learning, without target-domain reward feedback.

Multi-agent Systems (MAS) combine multiple model outputs to solve complex reasoning tasks. However, despite rapid growth of available open-source models, there is limited research on how to select optimal model candidates out of this massive pool. We systematically evaluate 8 model selection strategies (including model size, accuracy and answer diversity) across before-generation (routing) and after-generation (majority-voting, LLM-as-a-judge) MAS architectures on challenging scientific benchmarks. Our findings show a significant gap between theoretical oracle potential and actual performance: Expanding candidate pool sizes often degrades performance below that of the top performing base-model. We find that candidate selection within a single model family is the strategy that yields the best relative performance over a standalone model. These results demonstrate that adding arbitrary models to a heterogeneous MAS can introduce system instability, highlighting model selection as a critical design choice for multi-agent systems.

Capable open-weight models make local coding and reasoning attractive, but their context and execution state strain laptop memory. We present JustFit, an MLX-based inference runtime that combines KVExec for compressed KV execution, PhaseSwap for component residency, and StateTrans for state-preserving serving transitions. These mechanisms fuse reconstruction and coordinate just-in-time materialization and release, independently of model-weight quantization. In full-execution capacity tests on a 24 GiB M4 Pro MacBook running Qwen3.8-27B MXFP4, three independent runs complete 196,608 input and 16,384 output tokens, increasing completed single-request context from the mlx-vlm baseline's 30,720 positions to 212,992 (6.93x); a separate two-request run retains 229,376 positions in aggregate. In separate performance tests, a 32K-input, 64-output probe reaches 19.11 tokens/s, and a repeated 32K+6K workload has a median peak process footprint of 16,374 MiB. The integrated runtime answers 29 of 30 AIME 2026 problems correctly, showing how compact state and lifetime-aware execution expand local serving capacity while supporting extended generated reasoning.

Profound has raised a $180 million Series D at a $1.8 billion valuation, less than seven months after it raised a $96 million Series C.

Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking are our most advanced live dialogue models yet, built for natural conversation.