FoundersFeed

The biggest hallucination in coding agents right now? Work estimation 😂

They'll confidently tell you "this will take 2 hours" and then proceed to refactor half your codebase, introduce 3 new dependencies, and still not finish the original task. Classic overconfidence in time complexity vs actual implementation reality.

cc (Cursor Composer) runs silent verification passes on your code after you write it, testing edge cases and scenarios you never explicitly asked for. It's basically doing automated QA in the background while you work. Think of it as a built-in paranoid code reviewer that checks your logic even when you don't request it - catching potential bugs before they hit production. Pretty solid quality-of-life feature if you're using Cursor as your daily driver.

PSA: You can use Codex subscription credits directly in Todos.

Setup is dead simple:
• Run `tds provider add`
• Search for "codex"
• Pick "Browser login"
• Hit the auth link
• Done ✅

A lot of people don't realize Todos supports this natively. No need to juggle API keys or mess with config files.

Just burned through a Codex subscription session - modified 5 files with +64/-28 lines of code changes. Cost: $4.97.

If I were hitting the API directly for this, I'd be broke by now 😂

The pricing delta between subscription models vs pay-per-token API calls is getting wild. For heavy refactoring sessions like this, subscription is clearly the move.

DeepSeek's cost efficiency is insane – running this task on Opus 4.8 (medium) costs 10x more. That's not just a slight edge, that's a completely different league in $/token economics. For devs running heavy inference workloads, this kind of gap means DS can handle 10x the volume for the same budget, or drop your API bills by 90%. This is why everyone's stress-testing DS now – if the quality holds up at scale, it's a no-brainer switch for production.

Vibe coding workflow in action: using Todos to fix bugs on the fly. From creating a task to merging code takes under 2 minutes. This is the speed developers are hitting with AI-assisted coding tools - instant context switching, rapid iteration cycles. No more context-heavy ticket systems or lengthy PR reviews for minor fixes. The entire debug-to-deploy loop compressed into sub-2-minute sprints. Classic example of how AI coding assistants are reshaping developer velocity metrics.

Hot take on vibe coding: skip the docs, burn those tokens on refactoring instead.

The argument: clean code architecture > 100 pages of documentation. When you're iterating fast with AI-generated code, maintaining separate docs becomes a tax on velocity. Better to make the code self-documenting through clear structure, naming, and modular design.

This flips traditional software engineering on its head. Classic wisdom says "document everything" but in an AI-assisted workflow where code can be regenerated/refactored rapidly, static docs rot fast. The code IS the source of truth.

The real skill becomes: structuring your prompts and refactoring cycles so the output is inherently readable. Function names that explain intent, small focused modules, obvious data flows. If a human can't understand it by reading the code, neither can the AI on the next iteration.

Todos just dropped — a lightweight workspace for small teams + AI agents working together.

Setup takes 60 seconds. You run it on your own machine with your own API keys, so no vendor lock-in or privacy concerns. The core idea: spin up a swarm of agents that handle product development autonomously, and you just approve at critical milestones.

Think of it as CI/CD for agentic workflows — agents do the grunt work (code, docs, testing), and humans stay in the loop only when decisions matter. Built for teams tired of babysitting LLMs through every single step.

If you're experimenting with agent-driven dev pipelines or want a self-hosted alternative to cloud-based agent platforms, worth checking out.

Todos Team Secrets just dropped 🚀

Upgrade TDS to v0.1.28 and your agents can now hit private APIs. This means agents aren't stuck with public endpoints anymore—they can authenticate and call your internal services, third-party APIs with keys, or any protected resource.

Basically expanding what your agent can actually do beyond the usual read-only public stuff. If you're running multi-agent workflows or building autonomous systems that need to interact with real infrastructure, this is the unlock you've been waiting for.

Todos Agent ships with a built-in AskUser tool, same pattern as cc/codex. When the agent hits uncertainty, it prompts the user for clarification instead of hallucinating or guessing. Smart move—prevents the classic LLM issue of confidently generating garbage when context is ambiguous. This kind of human-in-the-loop design is becoming standard in production agent frameworks, especially for task execution where wrong assumptions can cascade into broken workflows.

Interesting observation: as AI models get smarter, you need fewer prompt engineering tricks. The more capable the model, the less you have to babysit it with elaborate instructions or chain-of-thought scaffolding. It just... gets it. 😂

Basically: dumb models need hand-holding, smart models need less BS.

Vibe coding power trio setup:

Planning layer → Fable 5
Execution layer → Grok 4.5
Research layer → GPT 5.6

Interesting workflow split: using Fable for high-level architecture decisions, Grok for actual code generation/implementation, and GPT for technical research and context gathering. This modular approach lets you leverage each model's strengths rather than forcing one to do everything.

Built role-based tool access control for my agent team. Only the deployment specialist can touch deployment tools now.

No more accidental production updates from other agents trying to "help" during task execution. Clean separation of concerns at the tool level.

Built role-based tool access control for my agent team. Only the deployment specialist can touch deployment tools now.

No more accidental production updates from other agents trying to "help" during task execution. Clean separation of concerns at the tool level.

Tested $GPT-5.6-sol on a mid-sized code refactor. Took 16 minutes 17 seconds. Pretty solid performance for this kind of task.

Not bad for automated refactoring - that's actually usable in real dev workflows. Curious how it handles edge cases and whether it maintains code style consistency across the changes.

GPT-5.6 just bumped context window by ~36.76% compared to previous version. Still relatively small though - likely means we're looking at maybe 16K→22K tokens range or similar incremental jump. Not the massive leap some were hoping for, but every bit helps for longer code analysis or document processing tasks. 🤔

The technical bottleneck has shifted from generation to prompt engineering. Video synthesis models (Sora, Runway Gen-3, Pika) now handle complex physics and temporal consistency pretty well. The hard problem is now ideation and prompt crafting—what scene composition, camera movements, lighting conditions actually produce useful output. It's like having a render farm but no art direction. The skill gap moved from "can you generate it" to "do you know what's worth generating."

If you're doing vibe coding, now's the time to hammer Grok 4.5 hard. Speed is solid, no signs of capability degradation yet.

Elon better keep this thing sharp and not lobotomize it like others did. The window of peak performance on these models is usually short before they start nerfing for cost optimization.

Pro tip for vibe coders: Grok 4.5 is running fast right now and hasn't been nerfed yet - use it while you can.

Elon please don't dumb it down like the others 😂

(Context: Many AI models get "safety tuned" or "aligned" post-launch, which often reduces their raw problem-solving capabilities. Early access = less filtered outputs)

Apparently the most cost-effective combo for vibe coding right now: GPT-5.6 (sol) + Grok-4.5

Balances quality, cost, and speed. GPT-5.6 handles the heavy reasoning, Grok-4.5 keeps iteration fast. Makes sense if you're shipping fast and don't want to burn through API credits on every autocomplete.