cobolcomesback

You do those things when you’re staying at a friend or family member’s place. But your friends and family aren’t charging you hundreds of dollars a night, plus a $250 cleaning fee.

It’s not just Airbnb the company that moved away from it, it’s the Airbnb hosts that moved away from it as soon as opportunists started becoming mega landlords renting out ikea-furnished party houses. It hasn’t been the original “person with a spare room for a weekend” Airbnb in a long, long time.

Those are all challenges that surely need to be overcome!

But they had _two months_ to figure out how to overcome them. For them to not have figured out a solution and have pre-built images ready to push as soon as the embargo lifted is ridiculous.

Nah, Chainguard should be praised for this. It was a different company that leaked the POC early, and then PHP, Wiz, OpenAI, and Calif sat on their hands doing nothing while the vuln was being actively exploited in the wild. Chainguard did the right thing by getting the fix out as quick as possible to stop the bleeding (or gushing flow of blood, really) rather than sitting on it so they could make a coordinated marketing post, like the others did.

GCP has some regions where all of the AZs are located in the same physical building. AFAIK they’ve stopped doing this in newer regions and have been migrating existing regions to have physically separate sites for each AZ.

AWS has always guaranteed that different AZs are physically separate sites, but that hasn’t always been the case with GCP.

AWS does not run “their control plane” out of us-east-1. Every single service has its own control plane independently hosted in its own region, with the exceptions of IAM, Route53, Cloudfront, and a small number of others like Global Accelerator.

IAM is usually the culprit when us-east-1 issues affect other regions, but Lambda runs independently in every region.

> Sure, and then tomorrow it's compaction, then worktrees, then an annoying buffer scrolling bug, image clipboard handling, etc etc. I don't buy for a minute that a one-liner prompt creates a perfect program because I know from experience it doesn't.

Every thing you listed works out of the box. Believe it or not, it’s true! And creating tools like this is only going to get easier as LLM models progress.

Your comment very much reminds me of comments from 2023 saying that LLMs would never be able to write working code, or be able to generate lifelike pictures, or be able to find security issues… but look around you, they’re doing all of those things already!

> I'm not sure I follow this line of thinking. Obviously, juggler is the result of an author who has spent a great deal amount of time thinking about an idea.

Awesome! That’s why I asked the author why they wanted attention. I want to hear more about those ideas behind juggler. What makes it special? Why does the author (who has tons of experience in software) think juggler is different? What were the obstacles that the author overcame while building it? That’s what I want to know, but these comments share none of that. Instead they just say “I also built a tool that does the same thing, I wish people would use my tool instead.” Why in the world would I?

I think you probably meant to write “underestimate”. But I think you are overestimating how difficult it is!

I’m not exaggerating when I say all I had to do was prompt “build me an LLM coding harness like Claude code, but make it so that I can resume sessions from OpenCode and Pi as well” and that’s it, and it was done within an hour.

Did it have warts? It’s been working nearly flawlessly, but there have been a couple of really minor things. One of the things I realized quickly was that it didn’t have a WebFetch tool built in, so I prompted “add a webfetch tool” and it did it within 5 minutes. Warts are trivially fixed.

The things that really differentiate software now are the ideas behind them, not the implementation. So if (like the OP) you’re looking at a tool and going “well mine does that too…”, so what? If it’s just a re-implementation of the same ideas, anyone can recreate it (and add on their own customizations to boot).

Personally, I am getting frustrated with the amount of people trying to pitch me their tool they made. Tools are a dime a dozen. Share with me your ideas, and I’ll share with you mine, but I’m not gonna adopt your tool.

Why do you want attention? Are you trying to sell this? Make money? Just want personal fame?

The reason there is so much noise in is area is because these tools are a dime a dozen and trivial to create with AI. I don’t think it realistic anymore to expect a rush of users and make money off of these. The supply is going to far outstrip the demand.

Last Saturday I asked Claude to create a Claude Code clone for me with a few customizations that I personally prefer. It did so in less than an hour, and works just as well as Claude Code. I’m not looking for attention on it - I’m happy I made it, and just myself as a user is gratification enough. I think this is the future of software development, and I think having expectations of attention from others just because you made another tool is setting yourself up for disappointment.

There are four different Pi plugins with the name “pi-mcp-adapter”, not including forks. Which one should they use for the fairest benchmark? And who should be the tie breaker when someone thinks they should use a different plugin? Now answer the same questions for all of the other plugins necessary for Pi to reach parity with Strands.

It’s like you didn’t even read my comment.

Base Pi doesn’t seem like an apt comparison here. Strands comes with MCP servers, subagents, and web fetch tools built in. To get those on Pi, you have to add addons, and then you get into the space of “which addons should be added to make it an apples-to-apples comparison? Which MCP addon do I use, the fastest one, or the most popular one?”. My guess is they chose OMP because it’s the defacto standard for answering those “which addons should be considered out of the box” questions for Pi.

The OP is not really a workflow manager, it’s a workspace manager that facilitates creating controlled environments where your skills can run. Everything you said is compatible with (and complementary of) the OP project.

With that said, I’ll somewhat disagree with you. I’ve been down the path you’re talking about and while it is incredibly flexible and powerful, it became too difficult to maintain, and too inconsistent between workflow runs, and a pretty hefty waste of tokens to use AI on things that could instead be handled by deterministic scripts. I ended up creating an orchestrator for myself that uses skills as the primary way to tell agents how to execute a step in a workflow, but also directly orchestrates running scripts and managing state in a deterministic way rather than leaving it all up to agents.

Ah, I see we’ve reached the point of “Don’t Look Up” where members of the public are actively mad that they’re being warned of the incoming meteor.

Amazon did exactly this with Alexa skills and it hasn’t exactly been a booming success. Turns out that all of the things you listed as examples are stuff that most consumers don’t actually want or will use.

“AI as a chat bot” has always been cool and fun in sci-fi movies, but it’s the least imaginative real world use for it. We need to move past it.

> $ Siri 'List the files in this folder'

Why in the world would I ever use this instead of just ‘ls’?

Considering you began this thread of discussion by cherry picking a single phrase from the GP and then contributing nothing to the discussion other than “you’re hopelessly naive”, perhaps you are the one who needs to read the HN posting guidelines.

If you post slop, you should expect slop replies.

It’s wild how many people think that these tech companies are capable of some major conspiracy where they’re gathering and tracking user data for some unknown-but-nefarious purposes… while simultaneously also ignoring or even rejecting the reality that this has already been happening, for years.

It’s like people claiming that Alexa is listening to their conversations to spy on them… when the reality is much worse. Every website you visit and app you download is spying on every single action you take, and they’ve gotten so good at it that they don’t need to listen to your conversations. They already know stuff about you without you even saying it out loud.

They infer your full name and address from the trackers baked into your OS, browser, phone, and every website you visit. They already have this. They’ve had it for years. They don’t need your ID for it.

Yea, “fun”, sure. I’m sure none of those billions of images were used to lie, mislead, or cheat anyone. I’m sure none of those images were used to create fake news narratives, used in advertisements to bait people into buying fake products, or create misleading mockups of rental properties. Definitely not.

Nevermind the fact that the examples in OpenAI’s blog post are examples of creating images that have no plausible purpose other than to fake stuff that didn’t happen.

But yea sure, it’s all just “for fun”. Sure.

Seriously, are these really the best examples they could come up with?

What is the point of having a fake picture of your dog in a costume? What’s the point of having a fake picture about being at a party?

The only use case I can think of for this is for someone who likes to make up stories and lie about what they’ve done. Is that really the target market?

This is just a rehash of the map already available at datacentermap.com (it’s literally cited as their source). Not sure what KXAN is adding here.

Also worth noting that a very loose definition of data center is used here that doesn’t fit the current colloquial usage of the term. One of the “data centers” listed in the middle of Austin, for example, is barely the size of a house and is actually just the switching infrastructure for a radio tower. Not every dot on this map is an energy-guzzling hyperscaler boogie man.

I’m a bit flabbergasted that an Anthropic engineer did reply to this issue, however their reply which they used Claude to write, and even includes some classic ‘Claude-isms’, claims they didn’t see any of the patterns being complained about.

Read the room Anthropic. Maybe don’t use AI to reply to a thread complaining about how AI output is hard to read.

I am so amused at the “HiLight” feature (lighting up colored LEDs for notifications so you can see them “at a glance” without looking at the screen). This is a feature that was on the Android phones of old (including the original G1) and I loved it, but it got phased out in favor of “always on displays”… which apparently now we want less of!

It’s funny how trends like this are a circle sometimes.

The route Qantas has stated they will use (at least for part of the year due to weather and winds) actually goes northeast over the Pacific Ocean, over the central US/Canada, over the North Pole, and then into the UK from the north. It doesn’t go over Asia at all.

OPs tool is not a sandboxer and isn’t like the other projects in your list. OP is an agent multiplexer, of which there also are many similar projects, but not a sandboxer.

The overwhelming majority of web app security issues, even the really complicated and sexy sounding ones, at their core are just “script kiddie style hacking”. Injection attacks and RCEs like this Jinja injection may seem “script kiddie” but they are a dime-a-dozen.

It is folly to assume that most companies don’t have similar weaknesses somewhere in their stacks. In the past, they were just sufficiently buried that they would get noticed and fixed before an attacker found them. With AI able to probe at thousands of these potential weaknesses per second… that’s not so true anymore.

There were almost daily posts in HNs front page about it 4 months ago when it first happened, as well as front page headlines on CNN and WSJ and other news outlets, though they quickly got drowned out by all of the other Iran war stories when the war first broke out.

This is just one of the popular ones, but there were also tens of posts about it with only 1-10 comments: https://news.ycombinator.com/item?id=47641788

Here’s another popular one: https://news.ycombinator.com/item?id=47632503

The reason it’s not really news this week is because both the Bahrain and the Dubai regions have been effectively non-operational for the last 4 months, so the IRGC saying they destroyed an already-destroyed region is less newsworthy.

The article in the OP is about a company in the EU (the clubs) not taking this seriously by outsourcing their stuff to an Irish company, who also is not taking it seriously. Hell, in the article, the CEO pretty much says “yea we ignored EU law, we will get fined, whatever”.

So I’m not sure the EU law is really working.

How so? The fact that China is going to launch a comparable model soon is the whole point of what’s happening now. Everyone knows there are going to be open models soon that have the same capabilities - Anthropic has literally said so. The restrictions now are to buy time to patch the security issues before those Chinese models are made available to the whole world.

Every one of the competitors capable of a similar model have been salivating for a long time at the idea of consensual data sharing. Anthropic just opened the door for everyone to do the same thing without having to deal with being the first to do so. My bet is that OpenAI etc’s next model will have these same requirements.

Ever since the Mythos announcement it’s been clear that we’re heading towards a future where SOTA models are no longer available to the average person, and not only cost more, but also require payment in the form of use case verification and data sharing. OpenAI’s 5.5-Cyber model requires the same, so it’s not just Anthropic.

We’re unhappy with this because we’ve all gotten used to being able to play with the new shiny model as soon as it’s available, but what I’m seeing in this thread about Anthropic being “stupid” is emotion-based wishful thinking.

That Claude support page says the exact same thing about AWS (“retained data stays in your AWS environment”). AWS’s docs say differently, though, so it seems one of them has incorrect documentation. I wouldn’t necessarily trust the Claude docs to be correct even regarding GCP until some of this is ironed out.

So why is OpenAI also releasing 5.5-Cyber in a private manner? Are they also out of compute?