What people are actually building, every day, summarized.
Sunday opened with @robinroy sharing a leaked benchmark claim that Gemini 4 Pro beats Astra on every benchmark at 6x lower cost with a 2M context window. Skepticism ran high (@samb…
Sunday's chat was dominated by orchestration philosophy: @jcartu, @tounano, and @thewildzeno pushed the 'smart orchestrator, dumb workers' pattern hard, arguing that Gemini 3.8 Fla…
Sunday's chat was heavy on model economics and workflow tooling. Builders debated whether to stack additional Claude 20x plans or add Astra 5x subscriptions, with @leewardbound rep…
September 18 was dominated by Jev, the new classifier/router model taking over X feeds. Builders shared real tests — @jasonakatiff clocked headless browser QA at 0.4s/call vs 8s on…
September 17 was dominated by two intersecting currents: the arrival of Jev/TypeSafe — a class of small, fast, sub-cent classifier models with confidence scores for structured deci…
September 16 was dominated by the launch of TypeSafe's Jev, a small, fast decision engine that returns confidence scores for classification, ranking, and boolean choices. Builders…
September 15 revolved around the Grok Bot Galaxy livestream, where xAI attempted to build a company in 3 days. Reactions split — the stream lagged, a data bot failed to create a Go…
September 14 was dominated by harness economics and orchestration architecture. @thewildzeno laid out a CEO/CTO/labourer model tiering framework — frontier models (Astra, Fable) fo…
September 13 was dominated by how builders are stacking, reviewing, and orchestrating coding models. Sentiment shifted hard against Opus — @jasonakatiff found 'a ton of issues' aud…
September 12 opened with builders waking up to another round of Codex weekly usage resets, sparking debate about Anthropic's strategy of pampering existing customers rather than op…
Builders spent September 11 pushing multi-model stacks, debating whether frontier models are being quietly degraded, and converging on new orchestration patterns. The day's biggest…
Supply shocks and model tuning dominated today. OpenAI officially paused new $200/mo Pro subscriptions and upgrades, sending builders scrambling for backup accounts, the hidden $80…
GPT Image 2.5 headlined the day, with builders wiring it into pipelines via Kie and the GPT sub and pairing it with Kling 3.0 Turbo for ad-quality video. Reference image handling i…
Big launch day: OpenAI shipped GPT Image 2.5, with builders wiring it into pipelines within minutes and reporting better reference-image positioning than v2, faster gens, and possi…
Astra's honeymoon ended on day three as builders reported loops, scope creep, and drift — the fix being a properly tuned harness and the emerging 'OMPHA' stack (OMP + Hindsight + A…
Sunday's dominant story was Astra's dramatic arc — early raves called it a 'generational leap' for frontend, UI/UX, and game design, with members canceling Claude subs to consolida…
Saturday was Astra's day. The model started appearing inside Codex for $200-tier ChatGPT users (logout/login required) and builders immediately called it a step-function upgrade ov…
Thursday belonged to Astra. OpenAI's new model (rumored GPT-6) began rolling out in the Codex model picker, and early testers led by @jcartu called it a bigger leap than Opus-to-Fa…
Tuesday's chat centered on model shakeups, ad platform pain, and infrastructure economics. Grok 4.6 got extended praise from @thewildzeno as a fast, pragmatic code-writing subagent…
Monday was dominated by Anthropic's surprise Fable 5.1 drop — builders reported dramatically faster inference, improved cache efficiency, and a less 'Claudish' tone, with most gett…
Saturday's conversation revolved around a shifting tooling stack. OMP emerged as the group's default coding harness, with @justingacina, @jcartu, @jarvisballer and others reporting…
Saturday's chat was dominated by video generation economics, with anonymous and @nowwatchthisdrive breaking down real per-second costs across Seedance 2.5 ($9 for 30s), Higgsfield…
Friday's conversation opened with a sharp debate on whether internal tools should be UI-first or agent-first. @tidemid argued that an agentic surface over your datastores makes tea…
Wednesday's chatter centered on model selection and harness maturity. A vocal thread pushed Grok as the main coder — fast, clean, and less preachy than Claude — with Jason Katiff,…
Wednesday's builder chatter split between a landmark performance milestone and a wave of practical stack blueprints. @jcartu cracked GLM 5.3 Flash on four 6k's, hitting 320 tokens/…
Wednesday's conversation centered on the Grok ecosystem and model choice. @jasonakatiff recommended running Grok alongside Codex for code reviews — Grok consistently surfaces more…
Apple's new Mac Studio with M5 anonymous/Ultra dominated Sunday's conversation, with 512GB configs landing at ~$17–20K and 4-unit clusters hitting ~4.8TB/s aggregate memory bandwid…
Sunday's discussion was anchored by @tounano's detailed walkthrough of a near-autonomous creative pipeline for native ads: a 100-ad decomposed library feeds headline generation, im…
Sunday's dominant theme was disciplined multi-model orchestration for coding. @jasonakatiff and GuruTime showed that Claude/Fable can directly invoke other CLIs headlessly (grok -p…
Saturday's chat was dominated by hands-on model orchestration experiments. anonymous led an in-depth Grok bot migration off Hermes — using Fable to write a handoff spec, then havin…