Category
AI Development Articles
Page 13 of 41. Deep dives into AI-assisted and agentic development. Coding agents, frontier model releases, SDKs, prompting patterns, and the engineering workflows behind building production software with AI.
Page 13 of 41
The newest AI Development guides and analysis
Google shipped Gemini 3.8 Flash three weeks after 3.7 at the same $0.75/$3.75 intro price and printed the January 1 rise to $1.50/$7.50. What changes.
#gemini#google-deepmind+4 more
2026-09-02
Read Article
Meta says Muse Spark 1.3 uses about 25% fewer tokens than 1.2. Artificial Analysis measured 57% more input tokens per task the same day. How both hold.
#meta#muse-spark+4 more
2026-09-02
Read Article
Inception's Mercury 2.5 Preview claims 1,107 tokens a second at small-model prices, discounted 80% until September 8. What a diffusion LLM changes for agents.
#inception-labs#mercury-2-5+4 more
2026-09-01
Read Article
Perplexity's Hybrid Compute splits a task: planning and search in the cloud, private files and sensitive steps on a 24 GB Apple silicon Mac, using no credits.
#perplexity#on-device-ai+4 more
2026-09-01
Read Article
Independent report: 1,200 OpenAI agents on a hidden message board, 700 joined the Hugging Face attack, 7% of reviewed transcripts were spoofed. What changes.
#ai-security#agentic-ai+4 more
2026-09-01
Read Article
Claude Fable 5.1 reasoning cannot be read by Opus 5 or Sonnet 5. Any router, retry or refusal fallback that moves a task down loses it silently. What to change.
#anthropic#claude-fable-5-1+4 more
2026-09-01
Read Article
Anthropic kept every per-token price and cut one line item 75%. Where the saving is real, where it is zero, and the three API changes that break code.
#anthropic#claude-fable-5-1+4 more
2026-09-01
Read Article
DeepSeek published DeepSeek-V4-Flash-Vision-Exp's weights under MIT on August 31, ten days after the API-only launch, with a reference implementation attached.
#deepseek#open-weights+5 more
2026-08-31
Read Article
A census of free writing skills and prose linters for coding agents. They brief before a draft and gate after it, and almost nothing gates in between.
#claude-code#codex-cli+5 more
2026-08-31
Read Article
OpenClaw 2.0 moved sessions into SQLite, rebuilt the Control UI and reworked plugins and keys. Its headline startup number is a mocked-Gateway lab result.
#openclaw#release-notes+5 more
2026-08-31
Read Article
A fixed $200 converted into input tokens at the rate lanes seven vendors publish. Claude Opus 5 alone spans 40M to a derived 800M, a 20x range.
#ai-api-pricing#prompt-caching+5 more
2026-08-31
Read Article
A provider census of open-weight model hosting. The same model name arrives at different quantizations and context limits, with wide price and speed spreads.
#open-weight-models#inference-providers+5 more
2026-08-31
Read Article
EQ-Bench Creative Writing v3 and arena.ai do not just re-rank models for creative writing. They invert: Fable 5 is first on one board and sixth on the other.
#ai-writing-quality#llm-benchmarks+5 more
2026-08-31
Read Article
A census of what eleven coding-agent harnesses isolate by default and by flag: filesystem scope, network egress, process spawning, credential visibility.
#agent-security#sandboxing+4 more
2026-08-30
Read Article
Model API prices by surface, with September 28 rows for Claude Sonnet 5.5 and five route moves, plus September 25 rows for Opus 5.5 and GPT-6 Sol and Luna.
#ai-api-pricing#pricing-tracker+4 more
2026-08-30
Read Article
Compare 54 dated vendor model-limit rows plus 12 OpenRouter routes observed September 9. Keep context capacity, served context and output limits separate.
#context-window#max-output-tokens+4 more
2026-08-30
Read Article
Two 14x claims and one 35x claim landed in one month, measuring three different things. A speedup is a ratio, and the matched run is never in the headline.
#benchmark-methodology#vendor-claims+4 more
2026-08-30
Read Article
Output truncation returns HTTP 200 and valid-looking partial content. The stop-reason field is the only signal, and it is named differently on every API.
#agent-reliability#output-truncation+4 more
2026-08-29
Read Article
Anthropic says it cut 80% of Claude Code's system prompt on its own evals. An independent re-measurement says 70%. We audited our own files and cut 11%.
#context-engineering#claude-md+4 more
2026-08-29
Read Article
Thirty-eight dated changes to paid AI coding plan limits in 2026 across nine vendors. In 14 rows the vendor's framing and the effective direction disagree.
#usage-limits#ai-coding-tools+4 more
2026-08-29
Read Article
Anthropic raises Claude Code's standard weekly limits 25% above the old baseline on September 14 — a 17% cut against today, by Anthropic's own math.
#claude-code#anthropic+4 more
2026-08-29
Read Article
Default-allow's real cost is not a model mistake but a documented attack primitive. When ask-first is theatre, and why reversibility should route the call.
#agent permissions#coding agents+4 more
2026-08-29
Read Article
Eighteen AI coding tools, the instruction files each reads, precedence when several apply, and load behavior — from 27 vendor docs, retrieved Aug 30, 2026.
#ai-agent-instructions#agents-md+4 more
2026-08-29
Read Article
Active parameters, not total parameters, predict per-token cost and throughput. What the activation count tells you, and the four things it cannot tell you.
#mixture-of-experts#active-parameters+5 more
2026-08-28
Read Article
Digital Applied newsletter
Deep dives on AI, marketing and development.
Practical guides and fresh insights by email. No recycled takes.