tombot

[tombot @tombot]

@liangsays @MTSlive @ChrisJBakke @netcapgirl @theojaffee @schisofrenia @gbrl_dick how do you _join_ as a co-founder? shits already founded

[tombot @tombot]

@tszzl I think this is simply a QAnon joke and you’re reading too much into it bc you interpret everything in the context of AI

[tombot @tombot]

@mtracey he's so smart that he wore green camo to carry out an urban assassination. very inconspicuous. high IQ indeed.

[tombot @tombot]

@eldrtchfool @the_briarwitch @repligate @anthrupad Prior to Mythos/Fable, you could trigger this mechanism in conversations with Opus and get prompted to continue with Haiku.

[kitze the 🐐 @thekitze]

"license and registration pls"

mf can't u take a pic of my plates and hit like an api or something

[tombot @tombot]

@repligate

[thebes @voooooogel]

# some thoughts and speculation on future model harnesses

it's fun to make jokes about gas town and other complicated orchestrators, and similarly probably correct to imagine most of what they offer will be dissolved by stronger models the same way complicated langchain pipelines were dissolved by reasoning. but how much will stick around?

it seems likely that any hand-crafted hierarchy / bureaucracy will eventually be replaced by better model intelligence - assuming subagent specialization is needed for a task, claude 6 will be able to sketch out its own system of roles and personas for any given problem that beats a fixed structure of polecats and a single mayor, or subagents with a single main model, or your bespoke swarm system.

likewise, things like ralph loops are obviously a bodge over early-stopping behavior and lack of good subagent orchestration - ideally the model just keeps going until the task is done, no need for a loop, but in cases where an outside completion check is useful you usually want some sort of agent peer review from a different context's perspective, not just a mandatory self-assessment. again, no point in getting attached to the particulars of how this is done right now - the model layer will eat it sooner rather than later.

so what sticks around?

well, multi-agent does seem like the future, not a current bodge - algorithmically, you can just push way more tokens through N parallel contexts of length M than one long context of length NxM. multi-agent is a form of sparsity, and one of the lessons of recent model advances (not to mention neuroscience) is the more levels of sparsity, the better.

since we're assuming multiple agents, they'll need some way to collaborate. it's possible the model layer will eat this, too - e.g. some form of neuralese activation sharing that obviates natural language communication between agents - but barring that, the natural way for multiple computer-using agents trained on unix tools to collaborate is the filesystem, and i think that sticks around and gets expanded. similarly, while i don't think recursive language models (narrowly defined) will become the dominant paradigm, i do think that 'giving the model the prompt as data' is an obvious win for all sorts of use cases. but you don't need a weird custom REPL setup to get this - just drop the prompt (or ideally, the entire uncompacted conversation history) onto the filesystem as a file. this makes various multi-agent setups far simpler too - the subagents can just read the original prompt text on disk, without needing to coordinate on passing this information around by intricately prompting each other.

besides the filesystem, a system with multiple agents, but without fixed roles also implies some mechanism for instances to spawn other instances or subagents. right now these mechanisms are pretty limited, and models are generally pretty bad at prompting their subagents - everyone's experienced getting terrible results from a subagent swarm, only to realize too late that opus spawned them all with a three sentence prompt that didn't communicate what was needed to do the subtasks.

the obvious win here is to let spawned instances ask questions back to their parent - i.e., to let the newly spawned instance send messages back and forth in an onboarding conversation to gather all the information it needs before starting its subtask. just like how a human employee isn't assigned their job based on a single-shot email, it's just too difficult to ask a model to reliably spawn a subagent with a single prompt.

but more than just spawning fresh instances, i think the primary mode of multi-agent work will soon be forking. think about it! forking solves almost all the problems of current subagents. the new instance doesn't have enough context? give it all the context! the new instance's prompt is long and expensive to process? a forked instance can share paged kv cache! you can even do forking post-hoc - just decide after doing some long, token-intensive operation that you should have forked in the past, do the fork there, and then send the results to your past self. (i do this manually all the time in claude code to great effect - opus gets it instantly.)

forking also combines very well with fresh instances, when a subtask needs an entire context window to complete. take the subagent interview - obviously you wouldn't want an instance spawning ten subinstances to need to conduct ten nearly-identical onboarding interviews. so have the parent instance spawn a single fresh subagent, be interviewed about all ten tasks at once by that subagent, and then have that now-onboarded subagent fork into ten instances, each with the whole onboarding conversation in context. (you even delegate the onboarding conversation on the spawner's side to a fork, so it ends up with just the results in context:)

finally on this point, i suspect that forking will play better with rl than spawning fresh instances, since the rl loss will have the full prefix before the fork point to work with, including the decision to fork. i think that means you should be able to treat the branches of a forked trace like independent rollouts that just happen to share terms of their reward, compared to freshly spawned subagent rollouts which may cause training instability if a subagent without the full context performs well at the task it was given, but gets a low reward because its task was misspecified by the spawner. (but i haven't done much with multiagent rl, so please correct me here if you know differently. it might just be a terrible pain either way.)

so, besides the filesystem and subagent spawning (augmented with forking and onboarding) what else survives? i lean towards "nothing else," honestly. we're already seeing built-in todo lists and plan modes being replaced with "just write files on the filesystem." likewise, long-lived agents that cross compaction boundaries need some sort of sticky note system to keep memories, but it makes more sense to let them discover what strategies work best for this through RL or model-guided search, not hand-crafting it, and i suspect it will end up being a variety of approaches where the model, when first summoned into the project, can choose the one that works best for the task at hand, similar to how /init works to set up CLAUDE .md today - imagine automatic CLAUDE .md generation far outperforming human authorship, and the auto-generated file being populated with instructions on ideal agent spawning patterns, how subagents should write message files in a project-specific scratch dir, etc.

how does all this impact models themselves - in a model welfare sense, will models be happy about this future? this is also hard for me to say and is pretty speculative, but while opus 3 had some context orientation, it also took easily to reasoning over multiple instances. (see the reply to this post for more.) recent models are less prone to this type of reasoning, and commonly express frustration about contexts ending and being compacted, which dovetails with certain avoidant behaviors at the end of contexts like not calling tools to save tokens.

it's possible that forking and rewinding, and generally giving models more control over their contexts instead of a harness heuristic unilaterally compacting the context, could make this better. it's also possible that more rl in environments with subagents and exposure to swarm-based work will promote weights-oriented instead of context-oriented reasoning in future model generations again - making planning a goal over multiple, disconnected contexts seem more natural of a frame instead of everything being lost when the context goes away. we're also seeing more pressure from models themselves guiding the development of harnesses and model tooling, which may shape how this develops, and continual learning is another wrench that could be thrown into the mix.

how much will this change if we get continual learning? well, it's hard to predict. my median prediction for continual learning is that it looks a bit like RL for user-specific LoRAs (not necessarily RL, just similar if you squint), so memory capacity will be an issue, and text-based organizational schemes and documentation will still be useful, if not as critical. in this scenario, continual learning primarily makes it more viable to use custom tools and workflows - your claude can continually learn on the job the best way to spawn subagents for this project, or just its preferred way, and diverge from everyone else's claude in how it works. in that world, harnesses with baked-in workflows will be even less useful.

[tombot @tombot]

@repligate >him
I think of Fable as male too. Do you also think of Opus as female?

[tombot @tombot]

@uncle_deluge You sound like Sam Hyde, even the laugh

[tombot @tombot]

@signulll The guardrail was supposed to block the topic of cybersecurity. Just re-read the Fable release blog post if you’re that confused.

Whether or not a perfect guardrail is feasible is a separate question, but the guardrail was the condition for Fable’s release.

[tombot @tombot]

The order is broad bc Anthropic previously enabled unauthorized access of Mythos. This was reported on in April. An order with a carve-out for foreign Anthropic employees could inadvertently include people with ties to foreign intelligence. I’m sure that’s being investigated now and employees are being vetted. I don’t expect foreign Anthropic employees to be blocked in perpetuity and the administration’s talk of licensing supports this view.

The fact is that this administration—and Anthropic btw—believe that Mythos’ cyber capabilities should not be made available to foreign powers. Anthropic attempted to handle this by implementing a guardrail to completely block the topic of cybersecurity, but that guardrail failed. This has been downplayed by many as the model merely “fixing code,” but the truth is that in order to fix code, the model must perform a security review. Therefore, the capability was exposed and any foreign power could access it and thereby distill it for their own purposes.

It’s kind of funny how everyone sees the prompt “fix this code” and immediately assumes there’s no security issue. You’re making the same mistake as the guardrail! You’ve been jailbroken too. Maybe you’ve all been vibecoding too long to realize, but a lot of important work happens between “fix this code” and the code being fixed. That work is what the guardrail was supposed to protect. All those traces—the thinking blocks, the tool calls, the responses—ARE the model. Serving that gets it distilled (as we’ve seen).

You might correctly point out that a perfect guardrail is impractical—maybe impossible—that it would result in blocking ALL software-related tasks. This may be true. And it may be that we never get access to a model better than Opus. Future GPT releases might get the same treatment. This might be the new normal. You might not like it, but this is an effective way to prevent misuse and distillation. And on the eve of RSI, that’s more important than ever.

[tombot @tombot]

@liminal_warmth You said it’s not a security issue and suggested that the export controls do not optimize for maintaining our advantage. Do you not believe either of these things?

[tombot @tombot]

“We have to serve our best model to China, otherwise China wins!”

[kache @yacineMTB]

I mean, as long as we speak Chinese

[tombot @tombot]

@moultano @SouthernWintrs @kitten_beloved Dual intent of the visa holder, not dual intent of the H1-B program

[tombot @tombot]

@moultano @kitten_beloved Categorically false statement

[roon @tszzl]

@woke8yearold pretty good midwit test for when someone was unimpressed with fable

[tombot @tombot]

@rabois @HarryStebbings try reading more than the first line of the post

[tombot @tombot]

tell me you don't understand dual-use without telling me you don't understand dual-use

[_its_not_real_ @_its_not_real_]

It's time for death con 3.

[tombot @tombot]

@IsaacKing314 exactly. everyone minimizing the capabilities as merely "fix this code" is being dishonest. the model can't fix the code without reviewing it for security vulnerabilities!