enraged_camel

I've been able to use Opus 5.5 and Fable 5.1 for defensive security audits without any issues. They cannot do offensive tasks like pen-testing but in a lot of cases that's not a big shortcoming.

>> But GP comment is not about how you feel about it. It is about where you direct that anger.

Don't worry, my anger is limitless and I can direct it at both corporations and corporations at the same time.

Based on some... rumors I've heard, this is their "Pro" offering. There is supposed to be an Ultra coming as well.

Sorry, that distinction makes zero sense. Either way you're paying a monthly opex cost, compared to it being your own hardware, in which case it would be a fixed capital expense. Which is what "your own infra" means. I worked in managed IT services for 15 years. Trust me, the terminology is important and words don't suddenly start to mean what you want them to mean.

>> "Running your own infra" also includes managed infra like Bedrock, Foundry, etc.

Managed infra is, by its very definition, not your own infra. It's infrastructure someone else sets up and manages for you.

Context is useful. The parent said: "it's a cheaper product that's almost as good or better in some cases"

The only open models that are "almost as good or better in some cases" require massive amounts of RAM. I posit that most people cannot afford a decked out Mac Studio, and therefore run the smaller "flash" variants on more normal devices. The issue is that those are nowhere near frontier-level in terms of capability.

>> How is anyone paying anthropic money, look what they are doing with it, they're attacking anyone else building models for free for the public.

That is not what they are doing. They are calling out specific providers who release powerful models without safeguards.

In addition, said providers are not "building models for free for the public." They are doing it to hamstring America's dominance in AI, primarily by undercutting the frontier labs.

>> And in the case of these open weight models: I can run it on my own infra and not give any data to anyone.

It's worth noting that the overwhelming majority of people who use Chinese models don't do this. Yes, it is nice to have the option, and there are US-based inference providers that claim to not send your data to China and maybe indeed don't, but in the grand scheme of things, we need to remember the adage that became popular during the social media era: if something is free (or, in this case, close to free), you are the product.

I think this is a flop. I watched the Livestream and the audience reaction at the end was very muted. You could always taste the "that's cute but can we move on already?" thoughts everyone had.

They are behind, hence the panic. On top of that, Sam has been trying to do another funding round, so he's desperate to make the company look good.

Opus 5.5 was a gut punch and my impression is OpenAI is still reeling.

You are painting half of the picture, perhaps on purpose? The other half is this: Opus 5.5 is significantly better than both Sol 6.1 and Astra, and with the newly increased limits across the board, it is quite difficult to run out (unless you're spamming agents at Max effort). So it is a much, much better deal than OpenAI's Pro 100.

This is a series of Ls for OpenAI that must have hit pretty hard. First Opus 5.5 crushes Astra 6, then Sonnet 5.5 comes in right behind that, and now they won't have an answer until at least November. Which of course gives Anthropic even more time to buff up Fable 5.5.

It depends entirely on its capabilities. If it is significantly smarter than Luna, which frankly is quite likely, then a lot of people won't mind paying more.

Sure, but what level of understanding do you actually need? Only you can be the judge of that obviously, but I posit that for the vast, perhaps even the overwhelming majority of any system's components, a high level understanding is sufficient. In fact that's how it already works: it is rarely the case that one person knows the team's entire codebase inside out. Instead different people specialize in different areas. With AI it's the same thing: you can achieve the level of understanding you want with the most important parts (either by having AI explain it to you, or architecting it and writing the code yourself), and delegate everything else to AI.

Subagents don't share context. But that's why delegating implementation to a subagent doesn't work well except for things that are truly mechanical in nature: the subagent needs to independently reason about the task it is given, and then the output will also be reasoned about by the main agent. So you end up wasting time and tokens.

Another amazing release. This, combined with Opus 5.5, puts OpenAI in an incredibly tough spot: it means Anthropic's both mid-tier models crush OpenAI's top-tier model in capability and are also faster and significantly cheaper.

If Astra 6.1 is released tomorrow during Dev Day it needs to leap-frog both, and considering 6.0 came out just three weeks ago I think that's unlikely. But even if that happens, Anthropic is still holding on to Fable 5.5, which rumor has it being prepared for release in the next few weeks.

OpenAI also has a more capable model codenamed 'Bel' but from what I hear that's a few months out at least.

It looks to me as if Anthropic not just killed but completely stole the momentum OpenAI had gained over the past few months. Even if Tibo showers people with resets it may not be enough to entice them back...

From TFA:

>> Claude Haiku 5.5, built for high-volume and cost-sensitive applications, will join the Claude 5.5 family in the coming weeks.

The way I think about it is that understanding is formed in a top-down fashion now, instead of bottom-up. You can look at a feature developed by AI from the outside, and keep peeling the layers and examining them (or having AI explain them to you) until you learn how it works. It's like reading a textbook. It is different than writing the code yourself, or practicing the topic of the textbook yourself. But if you invest the time, it can be just as effective.

The issue of course is that if you do invest the time, then you're no longer saving time by using AI. You're just spending it reading and trying to understand something you didn't write. And that can be unpleasant in its own way.

My hot take is that for parts of a system that can be considered its core, forming a deep understanding is almost always important, and so is knowing how the different business domains integrate and where the connection points are. For many others, a high level understanding is sufficient. The difference is that now, with AI, you can make that choice. Before, you had to write everything yourself, and for any sufficiently complex and long-lived system it became impossible to hold all of it in your head.

>> You're still doing 75% of what you did before.

No, I really am not. I'm doing maybe 10% of what I did before. The rest is filled up by other, usually higher order tasks like planning, product management and work orchestration.

Maybe you should define what "real intelligence" means, first.

In our context it would not be popular if it was not accurate and reliable. In fact we would lose customers pretty fast!

>> Chatbots are far WORSE than traditional UI for everything. If some product has a chatbot functions it's the first thing I disable, if it's not possible to disable it, I avoid the product.

The chatbot we added to our B2B product is by far the most popular addition we've made this year. Our users are not tech-savvy, they use a lot of apps everyday and don't want to have to learn and keep up with just another UI. So they like being able to type their wants and needs in plain language (or speak it into their phone, if they are in the field) and get a plain language response back with embedded images and charts.

YMMV of course.

Katsas and Rao, both Trump appointees, tag-teamed again. Trump automatically wins whenever they get a case.

Because, in the context he is describing, humans hunt and kill the wolves:

>> There's a reason why the last known wolf in Denmark didn't just wander off, but was shot dead in 1813.

>> When wolves get out of control, you shoot them.

>> I don't think users want to be overwhelmed with features, they want things that work nicely, and consistently.

In the case of Stripe, users want one thing, which is to receive correct answers to their questions nearly instantly. And from the sounds of it, that is indeed the case.

At this point we should be asking if there's anything or anyone OpenAI's agents didn't hack.

OpenAI's display of incompetence and negligence is absolutely stunning.

Wow, thanks for sharing your experience. Very insightful.

We tried Luna and it scored way lower in our evals. Muse also. We haven't had a chance to test others.

We use Haiku 4.5 inside our product. It continues to be absurdly capable for converting natural language to structured JSON based on a set of fairly complex business rules.

Ah, so you didn't read the article.

>> On our benchmarks, Claude Opus 5.5 leads in agentic coding, computer use, and knowledge work. That said, at these levels of capability we’ve found that benchmark margins have become a less reliable guide to real-world differences. In our own use, the gap between Opus 5.5 and Claude Fable 5.1 is narrower than these scores suggest.