4 ms·
For context: > Additionally, we’re introducing a new ultra mode that goes beyond the capabilities of a single agent by leveraging subagents to accelerate compl
by andai 4mo ago
For context:
> Additionally, we’re introducing a new ultra mode that goes beyond the capabilities of a single agent by leveraging subagents to accelerate complex work.
https://openai.com/index/previewing-gpt-5-6-sol/ https://openai.com/index/previewing-gpt-5-6-sol/
Can someone explain how this compares with Pro? I thought Pro was already something similar.
- ludamad 4mo agoI imagine this is something like Anthropic's dynamic workflows where a JS file is created to make a little AI harness on the spot
- andai 4mo agoWow, I hadn't heard of this! const audits = await pipeline(found.files, file => agent(`Audit ${file} for missing authentication checks.`, { label: file }), ) I asked Claude in the browser if it could do anything like that. It wrote a little frontend app that calls the Anthropic API (with fetch()), without including a key. I expected that to fail, but it worked! Apparently in the web chat (and also in Claude Code?[0] Though I haven't tried yet) they can call the Anthropic API and your subscription key gets auto-magicked into the requests somehow. Those are two separate things of course (aside from the key-injection) but I guess there's no reason it couldn't run completely in the front-end... hmm... [0] https://code.claude.com/docs/en/workflows https://code.claude.com/docs/en/workflows
- jaggederest 4mo agoI feel like it's not crazy to run Javascript in the browser... We've come so far I almost forgot where it all started.
- dlt713705 4mo agoPer latest available data, point of origin traces back approximately 13.8 billion years to a singular event designated "Big Bang". YMMV
- user_of_the_wek 4mo ago> In the beginning, the Universe was created. This had made many people very angry and has been widely regarded as a bad move.
- d4rkp4ttern 4mo agoMy understanding is that in the generated JS code, the subagents are invoked as headless Claude, equivalent to Claude -p, or the agents SDK.
- changoplatanero 4mo agoFor pro mode the agents worked independently and only when they all finished did a new agent take a look at everything to merge the work into a single response. The new thing involves subagents that have been trained to cooperatively pursue a task and are allowed to communicate with each other along the way.
- dools 4mo agoI tried a pro model out the other day and thought there must have been a bug in Pi’s cost calculations. But no, it’s absolutely fucking insane. Wasn’t even any better at the task.
- bombcar 4mo agoI really suspect that the models are basically the same below, it’s all in the prompt. The way I use them, surgically, they seem to perform about the same. Fable certainly hasn’t blow my socks off.
- giancarlostoro 4mo ago> Fable certainly hasn’t blow my socks off. Same. I suspect they'll get better at taking in terrible prompts over time though... Maybe that's what Fable does better, reminds me of Sora 2, it would take my crappy prompt and expound upon it. I told it once to generate a video of someone working at some company that changed its name, but the old name had historic relevance, it referred to the new company name without me telling it to, by virtue of me wanting a video of TODAY with a 90s icon.
- X-Istence 4mo agoWhere fable has blown me away is converting entire code bases and or refactoring across many different segments. It’s far more careful than opus and puts far more effort into testing and validating by default. Switching back to opus at work was a downgrade. Similar requests felt more clunky and needed far more hand holding.
- ilsubyeega 4mo agoi would believe this will be matched with something like orchestrator-focused model: https://news.ycombinator.com/item?id=48624782 https://news.ycombinator.com/item?id=48624782
- gbnwl 4mo agoHow is this any different than what we have already? We've had this ability for ages (6+ months, decades in the AI world), you can literally today easily prompt CC or Codex to use subagents to accomplish tasks and they'll do it well. My entire workflow is one top level orchestrator chat creating tickets to dispatch to subagents to implement, and other subagents to verify. Why is this being sold as a new thing? Have HN users never tried tried asking CC or Codex to use subagents?
- UnfitFootprint 4mo agoThe top comment on the thread explains this will involve subagent to subagent comms. To what effect I don’t know… I thought subagents were useful because they were explicitly single purpose and bound to a narrow context
- ai_fry_ur_brain 4mo agoBecause most people need complexity to be wrapped in a simple UI/UX. Most people just want the one-two button press and be on their way.
- tfehring 4mo agoI assume this is ~equivalent to ultracode in Claude Code, which can deploy a tree of hundreds of nested subagents and was just released experimentally 5 weeks ago IIRC.
- davedx 4mo agoI'd love to read more about how to use this workflow. What kind of top level instructions does this actually work with? Is there an article out there with some concrete examples of how to do this effectively?
- cyanydeez 4mo agoin opencode, if you directly referencee a subagent keyword or the name of a defined agent, it'll often spawn the agent. If you don't mention it directly, it's 50/50 whether any given request will invoke a subagent. The same with tools, skills, etc. No matter how smart these LLMs appear, they rarely do thinking as you expect. So basically: learn how the harnesses operate, and know the names of the tools they have.
- patentlyze 4mo agoPro also makes you ask it to use sub-agents instead of just doing it when useful. Hopefully, 5.6 will automatically spawn sub-agents without needing to ask.