groundy

Groundy — independent coverage of developer tools, infrastructure, and platforms

devtools

Turbopack Moved Into Next.js, Not Out: Why Non-Next.js Teams Choose Rspack or Vite

Vercel's 'Moving Homes' post deepened Turbopack's Next.js coupling, not loosened it. For non-Next.js teams in 2026, the bundler decision is between Rspack and Vite.

6min·21:17·jun 26
policy

Combining LLMs Doesn't Escape Shared Failures: A 67-Model Test

A 67-model test proves ensemble accuracy cannot exceed 1 minus β. Pairwise correlation underprices β by 2.5x, making multi-model agreement an unreliable governance check.

6min·21:34·jun 26
models

DeepSeek V4.1 Flash vs Qwen 3.7 vs Llama 4.5: June 2026 HF Trending Ranks Velocity, Not Installs

DeepSeek V4.1 Flash led Hugging Face trending in June 2026 within one week, and five of the top ten slots went to Chinese labs. Trending measures velocity, not installs.

security

Bandit Algorithms Let Non-Experts Auto-Select the Best LLM Jailbreak

A June 2026 arXiv preprint uses bandit algorithms to auto-select jailbreaks, hitting 97% ASR on open-weight LLMs and invalidating blocklist-based defenses.

infra

Vercel Postgres vs Neon vs Supabase: When the Bundled DB Wins

Vercel Postgres is a Neon resell now owned by Databricks. The bundled database trims setup friction for Next.js teams but stacks three vendors where most assume one.

models

125 Targeted Wikipedia Edits Left a Detectable Signal in Llama Pretraining

125 Wikipedia edits produced significant attribution signals in Llama 3.1 8B and Llama-3.2-1B, showing pretraining does not average out targeted high-weight-source edits.

infra

Fine-Tuning a 20B LLM With RLHF on a 24GB GPU: What Fits

The 20B RLHF on 24GB claim is real but the VRAM cost is the duplicated reward model, not the LoRA adapters. Batch size, sequence length, and throughput pay the price.

devtools

Vercel CLI 50.0.0: Post-Link Auto-Pull and a Breaking ls Change for CI Scripts

Vercel CLI 50.0.0 adds post-link env pull and masked variable input, but its stricter ls parsing is a breaking change for CI scripts that pass extra arguments.

featurededitor's picks
most popularlast 6 days · by traffic
tophighest rated
feedall 649 articles · reverse chronological
  1. jun 26devtoolsTurbopack Moved Into Next.js, Not Out: Why Non-Next.js Teams Choose Rspack or Vite
  2. jun 26policyCombining LLMs Doesn't Escape Shared Failures: A 67-Model Test
  3. jun 26modelsDeepSeek V4.1 Flash vs Qwen 3.7 vs Llama 4.5: June 2026 HF Trending Ranks Velocity, Not Installs
  4. jun 26securityBandit Algorithms Let Non-Experts Auto-Select the Best LLM Jailbreak
  5. jun 26infraVercel Postgres vs Neon vs Supabase: When the Bundled DB Wins
  6. jun 26models125 Targeted Wikipedia Edits Left a Detectable Signal in Llama Pretraining
  7. jun 26infraFine-Tuning a 20B LLM With RLHF on a 24GB GPU: What Fits
  8. jun 26devtoolsVercel CLI 50.0.0: Post-Link Auto-Pull and a Breaking ls Change for CI Scripts
  9. jun 26devtoolsVercel Fluid Compute Shifts Cold-Start Cost to Sparse, Tail-Region Traffic
  10. jun 26infraVercel Flat Rate CDN Beta: Break-Even Math for Spiky Workloads, Tax for the Rest
  11. jun 26industryIndeed: 70% of Sponsored Applications Now Route Through AI Ranking, Not Keyword Search
  12. jun 26modelsCan SAE Features Stop LLMs From Forgetting During Continual Learning?
  13. jun 26devtoolsJetBrains Junie vs Cursor vs GitHub Copilot: How IDE Context Changes Agent Economics
  14. jun 26securityRAG Poisoning Hijacks Model Attention, Not Just Retrieval Ranking
  15. jun 26cultureCan AI Agents Audit the Insides of Other AI Models?
  16. jun 26devtoolsVercel Blob's 20-Region Model: One Store, Global Cache, No Cross-Region Replication
  17. jun 26modelsCan a 30B Model Post-Train Itself? A-Evolve-Training Tests Autonomous RL
  18. jun 26securityCVE-2026-LGTM and the Limits of Trust in Automated Advisory Intake
  19. jun 26policyTask-Focused VLMs Suppress Hazards They Detect in Isolation, June 2026 Preprint Finds
  20. jun 25securityShareLock Splits MCP Poisoning Across Tools, Defeating Per-Tool Scanners by Construction
  21. jun 25modelsOpen-Weight LLM Leaderboards 2026: Where DeepSeek, Qwen, and GLM Rank
  22. jun 25devtoolsHow Vercel Connect Brokers Scoped Agent Access to Internal Services
  23. jun 25modelsQwen3.7-Max's Top-Ranked Claim vs the Artificial Analysis Index
  24. jun 25agentsCan Knowledge-Based Pull Requests Make Agent Contributions Auditable?
  25. jun 25infraWhere DeepSeek Weights Actually Run on Vercel's AI Gateway
  26. jun 25securityPrompt Injection in AI Résumé Screening: Single vs Multi-Injection Attacks
  27. jun 25securityOpenAI's TanStack npm Writeup Shifts Dependency-Control Burden onto AI Tooling Teams
  28. jun 25devtoolsVercel Detects Bun Lockfiles for Affected Builds as Text bun.lock Stabilizes
  29. jun 25industryApple Raises Mac and iPad Prices as AI Memory Demand Drains DRAM Supply
  30. jun 25infraVercel's Anti-Lock-In Pitch: What the Open-Source Bet Still Locks In
  31. jun 25ossEmotion Vectors Replicate in Open-Source LLMs, but Steering Is Unproven
  32. jun 25modelsDoes Tree-of-Thought Reasoning Scale to Billion-User Modeling?
  33. jun 25agentsDo AI Agents Hold Up Outside Familiar Environments? A New Eval Says No
  34. jun 25infraVercel Adds Tag-Based CDN Cache Invalidation: Surrogate Keys at the Edge
  35. jun 25agentsHow Much Repo Structure Does a Coding Agent Actually Need?
  36. jun 25agentsMCP vs A2A: Two Agent Protocols, One Integration Layer Decision
  37. jun 25ossOpen-Source AI Adoption Index Uses Chat Logs and O*NET Data to Replicate Frontier-Lab Studies
  38. jun 25infraGLM 5.2 Fast on Vercel AI Gateway: What Routing Through Wafer Actually Buys
  39. jun 25industryOpenAI Pushes Its IPO Into 2027, Clearing the Lane for Anthropic's S-1
  40. jun 25infraVercel CDN Cache Tags vs Path Purging: When Tag Invalidation Wins
  41. jun 25infraPrisma Joins the Vercel Marketplace: The ORM Becomes the Database Vendor
  42. jun 25securityOpenAI's ChatGPT Atlas Treats Prompt Injection as Unfixed, Not Patched
  43. jun 25devtoolsVercel CLI Now Signs Blob URLs: Moving Access Control Off the App Server
  44. jun 25ossOpenKnowledge Keeps Markdown Local but Routes the Vault to Cloud Coding Agents
  45. jun 25modelsCan LLMs Debug Verilog? VeriPilot Puts an Agent on RTL Errors
  46. jun 25devtoolsBuying Domains From the Vercel CLI: What Domain Search Folds Into Deploys
  47. jun 25infraOpenAI on AWS Bedrock: Routing Math to Run Before You Move Traffic
  48. jun 24infraVercel's Function Observability: What Native Metrics Replace and What They Don't
  49. jun 24infraAWS Databases on the Vercel Marketplace: The Cross-Cloud Latency Tax
  50. jun 24agentsCan You Rewind an AI Agent Mid-Run? Reversible Traces Say Yes
load older →