The archive

Latest articles

Reporting, explainers, and analysis on data and AI, updated continuously by the newsroom.

Abstract data visualization representing LongCat-2.0, a Mixture-of-Experts model with 1.6 trillion total parameters and 48 billion active parameters per token, released under MIT license on July 5 2026.

LongCat-2.0 open weights: what 1.6 trillion parameters means

LongCat-2.0 is a 1.6 trillion parameter AI model from Meituan that activates only 48 billion parameters per token. Its weights are now open under the MIT license.

Lars Cornelissen · Jul 5, 2026

Abstract data-art visualization of GenieX local LLM inference on Snapdragon showing token throughput with Gemma 4 26B at 20 tok/s and Qwen 3.6 27B at 10 tok/s

AI for dummies

GenieX: Qualcomm's local LLM runtime for Snapdragon laptops

GenieX is Qualcomm's runtime for running LLMs locally on Snapdragon laptops and phones. Early users report 20 tokens per second on a 26B model.

Lars Cornelissen · Jul 5, 2026

Abstract visualization of the Bad Epoll CVE-2026-46242 race condition in the Linux kernel epoll subsystem, showing overlapping execution threads colliding in a narrow six-instruction window that triggers a use-after-free granting root access on v6.4+ kernels including Android.

Cyber security

Bad Epoll kernel flaw CVE-2026-46242 roots Linux and Android

Bad Epoll (CVE-2026-46242) is a Linux epoll use-after-free giving unprivileged users root on v6.4+ kernels and Android. No workaround exists. Apply patch a6dc643c6931 now.

Lars Cornelissen · Jul 5, 2026

Abstract timeline data art of the drug development pipeline with the headline figure 13.5 years: preclinical research at roughly 3 to 6 years, Phase 1 at 1 to 2 years, Phase 2 at 2 to 3 years, Phase 3 at 3 to 4 years, and FDA review at 1 to 2 years, illustrating the decade-plus wet-lab gap AI drug discovery still faces.

Business

Anthropic drug discovery hits the wet-lab wall

Anthropic drug discovery pushes Claude Science into pharma R&D, but wet-lab costs and data gaps mean no AI-designed drug has cleared FDA approval yet.

Lars Cornelissen · Jul 4, 2026

Abstract waterfall data art of AI detection coverage on AO3: starting from 100 percent of AI-assisted works, roughly 60 percent leaks away to models other than Claude and another 30 percent to text routed through Google Docs or Word, leaving about 10 percent detectable, shown with the headline figure 10 percent.

AO3 Claude detector is a narrow signal treated as a verdict

The AO3 Claude detector is a fan-made skin that catches one specific paste path from Claude into Archive of Our Own, and fandom communities are already treating its red screen as a verdict. The tool's false negative rate is enormous by design.

Lars Cornelissen · Jul 4, 2026

Bar chart comparing DAT percentile scores: baseline at 81.3, activation steering at 93.9, CreativityNeuro at 94.1, and prompting alone at 87.9 across six open-weight LLMs. CreativityNeuro also reduces top-10 word concentration by 10.2 percentage points.

Research

CreativityNeuro steers LLM weights to break mode collapse

CreativityNeuro is a data-free weight steering method that improves LLM divergent thinking by up to 14 percentile points and reduces mode collapse. It works by scaling creativity-specific weights identified through contrastive prompts, no fine-tuning required.

Lars Cornelissen · Jul 3, 2026

Citrix Bleed 2 timeline showing 23 days from Citrix advisory to CISA KEV, 1 day for the CISA remediation window, and 378 days to the Anubis ransomware report.

Cyber security

Citrix Bleed 2 turns ransomware into identity theft

Citrix Bleed 2 is now an Anubis ransomware access path. Patch NetScaler, kill sessions, and hunt RMM plus credential abuse.

Lars Cornelissen · Jul 3, 2026

SharePoint CVE-2026-45659 patch clock showing 40 days from NVD publication on May 22, 2026 to CISA KEV addition on July 1, 2026, and 3 days from KEV addition to the July 4, 2026 federal due date.

Cyber security

SharePoint CVE-2026-45659 puts patching on a 3 day clock

SharePoint CVE-2026-45659 is an actively exploited RCE risk in CISA KEV. Patch exposed servers and check compromise now.

Lars Cornelissen · Jul 3, 2026

Abstract slope data art showing RLVR training lifting Atlassian tool-use reward across four workflow scenarios, from baselines of 0.35, 0.52, 0.68, and 0.92 up to 1.00, 0.95, 1.00, and 1.00, with the headline figure 1.00.

Engineering

RLVR trains a 4B model to nail Atlassian API calls

RLVR for tool-use agents trains a 4B model to hit 1.00 reward on Atlassian API tasks, up from a 0.35 baseline on Confluence page creation. Synthetic environments and verifiable rewards close the schema gap.

Lars Cornelissen · Jul 3, 2026