The Groundy archive · Page 36 of 43
Browse Groundy's complete archive of 1027 articles on AI, developer tools and infrastructure. Page 36 of 43.
841–864 of 1027 articles · Newest first
Fable 5 ships broad biology and chemistry classifiers that route flagged prompts to Opus 4.8. Here is what that fallback means for biotech teams and long-running workflows.
Claude Fable 5 is free on subscription plans through June 22. From June 23 it draws usage credits at $10/$50 per million tokens. Here is what that means for team budgets.
Claude Fable 5 ships with distillation protection to prevent capability extraction. A first-principles look at what it is, how it works, and why API consumers should care.
POISE achieves 89.3% attack success on codex+gpt-5.2 by placing malicious instructions where agents naturally execute them, making static content scanners effectively blind.
Cloudflare claims a 15x bot surge using a classifier that flags privacy browsers as bots. Audit your own logs before trusting the numbers behind Pay-Per-Crawl.
OpenAI's 3M daily compensation queries push ChatGPT into salary benchmarking, but the model lacks the proprietary employer panels behind Radford and Mercer's moat.
An arXiv preprint shows GPU inference outputs can be reproduced bit-for-bit across hardware, giving auditors a forensic trail to verify which model produced a given output.
Two June 2026 preprints show VLA robot policies already compute safety-relevant signals at inference, enabling real-time collision monitors with no retraining.
Bloomberg's Pomona agent limits diffs to 10 lines and merged 88% of PRs in production, proving small, bounded edits earn reviewer trust faster than large autonomous refactors.
PROVE and AgentTrust show learned policies beat hand-tuned rules for gating AI agent tool calls, but the gains depend on calibration that neither paper measures.
Splitting a disallowed action into benign tool calls bypasses per-call safety filters in LLM agents, lifting jailbreak success by 28 percentage points over current baselines.
ICML 2026 research finds o3 achieves only 17% of optimal cooperation while weaker o3-mini hits 50%, proving model capability does not predict multi-agent coordination.
Three papers show safety alignment can be extracted as a portable adapter and reapplied to fine-tuned models, replacing per-model alignment with one adapter per model family.
Bending Spoons' F-1 shows $1.31B revenue and 95% retention across 50+ brands. Evernote's 24/100 churn score suggests the roll-up model harvests more than it compounds.
OpenAI's GPT-5 API features map directly to Cursor's agent loop, revealing a co-design relationship. Coding teams must now evaluate editor-model pairs, not editors alone.
DuckDB queries Parquet files on Hugging Face Hub over HTTPS without downloading them first, turning dataset triage from a multi-gigabyte commitment into a LIMIT 100 query.
A 2026 arXiv preprint argues RLHF's single reward signal destroys the reasoning behind human disagreement, proposing machine theory-of-mind as an alignment foundation.
A June 2026 arXiv analysis traces an AI program through the DoD Software Acquisition Pathway, finding no milestones for model re-validation or data provenance.
The TRAP benchmark finds 13 to 43 percent of web agent tasks can be redirected by persuasive page content, exposing a blind spot in current instruction-hierarchy defenses.
OpenAI's URL provenance filter concedes content inspection is intractable. Agents that mix sensitive data with web access face a structural exfiltration risk.
A June 2026 paper frames the AI agent benchmark gap as a sim-to-real problem, giving eval teams a four-part MDP checklist to challenge vendor claims before live deployment.
Vercel's shift from frontend hosting to agent infrastructure is backed by real products and a $9.3B valuation. Whether per-token billing beats per-seat SaaS remains unproven.
A new framework decomposes patent value into per-feature Shapley credits, but courts have not ruled on whether model output replaces expert testimony in damages and M&A.
A June 2026 paper proposes Data Flow Control, moving agent data safety from prompt-level guardrails to deterministic, auditable SQL query policies enforced outside the model.