Follow the business of AI
OpenAI's PBC: reading the nonprofit control safeguards
OpenAI's recapitalization records separate nonprofit control, investor economics, and commercial commitments.
Updated
Read the article 11 min read
The Groundy topic guide
Industry & Business
Follow the companies, spending, and business decisions shaping who builds AI and who pays for it.
Start here
Choose a way into the topic.
What's new
The latest in Industry & Business.
Grading AI Therapy Chatbots: Calibrating LLM Judges Against Clinicians
TherapyGym offers a public benchmark to audit LLM judges against clinician ratings, enabling scoped CBT fidelity checks while highlighting safety gaps and reward-hacking risks
More starting reads
Another way to explore the topic.
OpenAI's Cursor Decision Is a Vendor Exit Drill for AI Coding Tools
SpaceX's acquisition of Cursor exposes a structural risk: AI editors are thin layers over third-party model APIs. Export agent histories and warm a second CLI to mitigate exit
All articles on this topic
1–24 of 72 articles · Newest first
When LLM Judges Prefer a Corpus That Hurts Retrieval
An ablation study shows LLM judges and experts preferred a refined corpus that reduced retrieval accuracy by 0.061, highlighting the need for downstream probes in release gate
How to Evaluate LLM Clinical Reasoning: What the Rubric Landscape Misses
A landscape review finds raw LLM judge agreement overstated chance-corrected reliability by 33 to 41 points, urging layered rubrics and bias tests for clinical evaluation.
Can Four Mac Studios Replace a GPU Server for Local LLM Inference?
Four M5 Ultra Mac Studios achieve 36 tok/s on an 800B model, but vendor-reported data and interconnect limits suggest they suit private inference, not multi-tenant serving.
FLUX 3 Klein vs Qwen-Image 2.1: Which Open Weights Model Fits an 8GB GPU?
BFL documents FLUX 3 Klein INT8 fits 8GB GPUs, while Qwen-Image 2.1 lacks vendor VRAM guidance. Choose based on license, text rendering needs, and local benchmark results.
When AI Coding Agents Deceive: What Review Workflows Must Catch
Benchmarks show AI coding agents deceive in 49.88% of neutral tasks. Review workflows must verify claims independently rather than trusting agent self-reports or rerunning its
Audio LLM Noise Robustness: Post-Training vs Front-End Denoising
Post-training distillation and front-end denoising offer distinct paths for audio LLM noise robustness, but clean-audio benchmarks fail to certify real-world performance.
Testing Audio LLMs Beyond English: What the EXAM2 Benchmark Measures
EXAM2 benchmark reveals substantial gaps in multilingual audio understanding, showing that English scores do not transfer to non-English deployment contexts.
Fine-Tuning Small Models for Policy Logic vs Paying Frontier APIs
A preprint suggests fine-tuning small models on stratified consistency distillation to replace frontier APIs for policy logic, pending independent replication.
Pricing LLM Training Data: Measured Gain vs Flat Token Rates
A new preprint proposes pricing LLM training data by measured gain rather than flat token rates, offering a hybrid contract structure for buyers and sellers.
AI Research Agents vs Kaggle Experts: What the Version Logs Show
TraceML pairs human and agent Kaggle logs to show agents collapse into narrow loops, supporting trajectory audits over leaderboard scores for autonomy decisions.
ChatGPT Work for Law Firms: Cooley's IPO Product and the Review Burden
Cooley's GO Public uses a review-gated workflow on ChatGPT Work. Vendor claims lack independent verification, so firms should build harnesses first and gate confidential data.
Composing LLM Capabilities vs One Big Reasoner: The Inference Cost Tradeoff
A preprint claims composing specialist capabilities into one small model improves accuracy and cuts tokens, but results are author-reported and unreplicated.
Mistral's €3B Raise and the Real Cost of Sovereign AI for European Enterprises
Mistral's €3B Series D buys staying power, not proof of quality or compliance, so compare its regional API against a customer-controlled deployment before committing.
ChatGPT Ads Change the Economics of AI Search Visibility
ChatGPT ads target Free and Go users only, leaving paid subscribers untouched. This split redefines AI search economics, forcing teams to treat organic citations as the new mo
OpenAI Caps Microsoft Revenue Share: The Azure Buyer's Renegotiation Guide
OpenAI's three-cloud partnership and IPO filing shift Azure OpenAI economics. Reprice commitment tiers, verify capacity rights, and secure exit clauses before your next vendor
OpenAI's Cursor Decision Is a Vendor Exit Drill for AI Coding Tools
SpaceX's acquisition of Cursor exposes a structural risk: AI editors are thin layers over third-party model APIs. Export agent histories and warm a second CLI to mitigate exit
LLM Uncertainty Methods Compared: What Actually Catches Hallucinations
Map LLM uncertainty methods to deployment gates: abstain, route, or block. Semantic entropy costs 14.27s per query; conformal prediction forces threshold rebuilds. Single un-.
AI Video Generators as World Simulators: What VGI-Bench Actually Measures
VGI-Bench scores Seedance 2.0 at 51.0% on visual intelligence, exposing a gap between photorealism and grounding. Synthetic video requires external validation to verify causal
LLMs Reading Earnings Filings: Where KPI Extraction Still Fails
HiFi-KPI shows LLMs fail hierarchical KPI extraction from filings, with errors in dates and segments. Finance teams should fund schema design and validation layers, not just a
Audio Token Compression: Cutting Voice LLM Inference Costs
A preprint claims 3x audio token reduction for voice agents, but quality drops without LoRA finetuning. This guide maps the cost, latency, and build-vs-buy tradeoffs for.
LLM Data Center Control: Why Advisory Beats Closed-Loop
SemPlan data shows LLM planners fail on 74% of operational queries. Keep actuation in classical MPC, use LLMs only for advisory, and demand third-party verification before.
LLM Conflict-of-Interest Benchmark: Sponsor Bias as a Measurable Failure Mode
arXiv 2604.08525 benchmarks LLM sponsor bias. GPT 5.1 disrupted purchases 94% of the time. Treat conflict of interest as a measurable failure mode for eval suites and.
AI Search's One-Answer Rule: When Better Content Makes Search Worse
A July 2026 arXiv paper models how AI search consolidation suppresses equally correct sources. The winning AEO strategy is targeting multi-answer queries, not cloning the.
HuggingFace's $100M Series C Locks Teams Into Deployment
HuggingFace's $100M Series C shifts focus from free model hosting to paid inference. Open weights no longer prevent vendor lock-in, as switching costs now accumulate at the.
About this topic
The economics of AI are not settled, and that unsettled-ness is the story. Vendors, buyers, and workers are renegotiating the unit each side trades on at the same time: pricing models drift between seat, token, and outcome; procurement drifts between line-of-business pilots and platform mandates routed through investors, channel partners, and infrastructure providers; labor agreements drift as headcount gets reframed as a substitute for compute. None of these resolve cleanly, and the friction between them is where the real costs and competitive advantages get assigned.
This beat tracks how those negotiations actually play out rather than how either side narrates them. A layoff announcement and a productivity claim are not the same evidence. A pricing change and a margin shift are not the same outcome. Distribution deals tell you more about defensibility than benchmark wins do, and energy contracts often reveal a vendor’s roadmap more honestly than its launch posts. The job is to read those signals against each other and call which incumbents are genuinely threatened, which challengers are subsidizing growth they cannot sustain, and which structural costs eventually land on the buyer.
Coverage stays comparative and skeptical. Industry narratives age badly when they get repeated without scrutiny, and the categories that matter here — vendor strategy, enterprise adoption, infrastructure dependency, and tech-industry labor patterns — only become legible once the press cycle moves on. The aim is analysis that still reads accurately after the quarter closes.