programd

So there are a lot of ad-hoc solutions everybody is using for agent isolation and Drop looks well thought out. But the obvious question is what do the pros use? What does Anthropic, OpenAI, SpaceXAI, Google and Amazon do in house to deal with the isolated agent environment problem?

Based on past experience I'm guessing everybody is using Kubernetes, but what are the details? Stripped down VMs on Kubernetes? Hardened containers on Kubernetes? Can the rest of us spin up something comparable on K3s?

Anybody knows the answers and can share?

The star map screenshot looks very similar to star map in "Remnants of the Precursors" [1], which is a modernization of 1993 "Master of Orion" [2]. Convergent evolution, or was 1995 Stars! an homage/clone of Orion?

1. https://rayfowler.itch.io/remnants-of-the-precursors

2. https://store.steampowered.com/app/410970/Master_of_Orion_1/

If you add all that you will have recreated “Silent Hunter 4: Wolves of the Pacific”

https://store.steampowered.com/app/15200/Silent_Hunter_Wolve...

The hardware is a helmet with a buch of sensors and IR LEDs on the inside. There's a writeup of the process and a picture of the outside here:

https://condu.it/thought/10k-hours

Their data acquisition routine was driven by Claude and they were recruiting random people off Craigslist pretty aggressively a few months back. I know a couple of folks who responded out of curiosity and described a very janky setup in the rented basements of a couple of residential houses in SF. From what I heard about the setup the data they got must have been incredibly noisy.

To the people who are sceptical here, I wouldn't count them out yet. They almost certainly can get funding with what they captured, which would get them better hardware. And given the state of brain scanning - people have captured images from scanned brain activity some years ago as I recall - I'd say they have a good shot. But yeah, it's not telepathy.

Previous HN discussion:

https://news.ycombinator.com/item?id=46195109

Also worth reading is the blog post by the Cloudflare CIO Sam Rhea on how they are actually using it internally, with some examples.

https://blog.cloudflare.com/how-we-use-ai-with-cloudflare-os...

Cloudpunk and its sequel City of Ghosts should be on the list. Very laid back games heavy on charming atmosphere with a great soundtrack. The voxel city design is excellent.

https://store.steampowered.com/app/746850/Cloudpunk/

The feedback from the VC crowd is that the historical 5 year startup timeline is down to 18 months these days. That means finding PMF, funding rounds, hiring and scaling to $25 mil/year. That's the VC playbook changer which has them all worried about businesses which don't go all in for AI.

Define easily. There is an approach that apparently works and is based on spectral analysis of the images.

https://github.com/aloshdenny/reverse-SynthID

His RISKS Digest should be required reading for anybody studying computer and software engineering. RIP Peter.

It's partly the weight that made the 11" MacBook Air great:

  MacBook Air: 1.08 kg (2.38 lbs)
  MacBook Neo: 1.23 kg (2.70 lbs)

Thomas Wake: "Hark Triton, hark! Bellow, bid our father the Sea King rise from the depths full foul in his fury! Black waves teeming with salt foam to smother this young mouth with pungent slime, to choke ye, engorging your organs til' ye turn blue and bloated with bilge and brine and can scream no more - only when he, crowned in cockle shells with slitherin' tentacle tail and steaming beard take up his fell be-finned arm, his coral-tine trident screeches banshee-like in the tempest and plunges right through yer gullet, bursting ye - a bulging bladder no more, but a blasted bloody film now and nothing for the harpies and the souls of dead sailors to peck and claw and feed upon only to be lapped up and swallowed by the infinite waters of the Dread Emperor himself - forgotten to any man, to any time, forgotten to any god or devil, forgotten even to the sea, for any stuff for part of Winslow, even any scantling of your soul is Winslow no more, but is now itself the sea!"

Ephraim Winslow: "Alright, have it your way. I like your cookin'."

See also: Clarkson's Farm [0], for some of the messy reality of running an actual modern farm in England (though edited for entertainment value). I suspect the current AIs are not quite up to doing this - but I firmly beleive it's only a matter of time.

[0]https://www.imdb.com/title/tt10541088/

You could mod the game with more varied events, which were of course AI generated to begin with. Bit of an inception scenario where AI plays an AI modded game.

The other option is to have an AI play another AI which is working as an antagonist, trying to make the player fail. More global plagues! More scheming underlings! More questionable choices for relaxation! Bit of an arms race there.

Honestly I prefer Crusader Kings II if for no other reason that the UI is just so brilliantly insanely obtuse while also being very good looking.

I'll go further. Ignore the Perl specific bits and Conway's "Perl Best Practices" is one of the best general programming books ever written.

It has so many great pieces of advice that apply to any programming task, everything from naming variables, to testing, error handling, code organization, documentation, etc, etc. Ultimately, for timeless advice on programming as a profession the language is immaterial.

Global rollout of security code on a timeframe of seconds is part of Cloudflare's value proposition.

In this case they got unlucky with an incident before they finished work on planned changes from the last incident.

My armchair is equally comfy, and I have an actual paper to point to:

Jaxley: Differentiable simulation enables large-scale training of detailed biophysical models of neural dynamics [1]

They basically created sofware to simulate real neurons and ran some realistic models to replicate typical AI learning tasks:

"The model had nine different channels in the apical and basal dendrite, the soma, and the axon [39], with a total of 19 free parameters, including maximal channel conductances and dynamics of the calcium pumps."

So yeah, real neurons are a bit more complex then ReLU or Sigmoid.

[1] https://www.biorxiv.org/content/10.1101/2024.08.21.608979v2....

SEEKING WORK | Remote | Onsite San Francisco Bay Area |

Versatile back end code and infrastructure development services. Typical deliverables are Dockerized microservices specified by OpenAPI, written in Go or Python, with a REST/JSON API and backed by PostgresDB or MySQL. CI/CD pipelines included and deplyment on any of the major cloud vendors. And of course your custom business logic.

If you have more complex needs, e.g. project or team management please reach out as I have extensive career experience in all aspects of of technology development and management in startups and large enterprises. Need a contract manager or CTO perhaps?

At minimum I can help you develop your back end infrastructure from the ground up. Basically I offer development of back-end componenets you can slot into your Kubernetes or Docker environment from day one.

On the business side, you get a fractional developer for a no-haggling fixed monthly rate, corp-to-corp billing, long term support for your code, careful vetting of dependencies for licensing and security, and a professional approach to your technical needs.

Contact: contact@nulladmin.com

Website: https://nulladmin.com/

LinkedIn: https://www.linkedin.com/in/abnulladmin/

  Location: SF Bay Area
  Remote: Yes - or hybrid/in-office, I'm flexible
  Willing to relocate: No
  Technologies: Go, C/C++, Python, Docker, Kubernetes, PostgreSQL/MySQL, ML/LLM, and many others
  Résumé/CV: https://linkedin.com/in/abnulladmin
  Email: contact@nulladmin.com

I'm an engineer with many years of experience in Silicon Valley, working in management roles, engineering, and consulting. My expertise is mostly in back end systems, working in the modern Go/Docker/Kubernetes ecosystem. In addition to working directly on large and complex code projects I also managed and grew engineering teams at various companies.

If you need help with specific technical projects, or are looking for management roles please feel free to reach out. I'm very familiar with both the unique challenges of scaling up startups as well as working to innovate in large enterprise companies.

Specifically if you need help productizing and scaling out various AI/Machine Learning/LLM systems this would be a good fit. I know there is quite a bit of demand in this area and a limited supply of qualified candidates.

I'll second that recommendation. One of the more interesting videos is AI creating utopian and dystopian worlds in Minecraft

https://www.youtube.com/watch?v=FCnQvdypW_I

Invokes shades of Iain M. Banks' "Surface Detail" Culture novel, where virtual Hells are a major plot point.

More discussion on Ubuntu forums

https://discourse.ubuntu.com/t/problem-with-updates-error-fa...

The worst part is that they marked the incident as resolved after 32 minutes, but didn't mention that mirrors for the packages on security.ubuntu.com have huge queues. OK fine, we can wait until the mirrors sync and you can choose another mirror to do your update - eventually. You can also work around this while updating Ubuntu 24.04 by manually installing the deb file it wants.

But wait, there's more! You can't install new instances of Ubuntu 24.04.2 because the installer connects to security.ubuntu.com by default (probably for good reasons) and will bail out while formatting and writing the disk when it gets 500 Internal Server Error from security.ubuntu.com for a specific deb file. There's no option around it that I'm aware of if you're doing an install connected to the network. I'm told that things should work if you try to install without networking connected. But that's not working for me, possibly due to some drivers it needs to pull for my hardware that are not in the default installer.

Ran into this while trying to install a fresh instance on an old mac hardware.

All in all, not a good look for Canonical, especially given how long this is taking to resolve and the lack of any status indication that this is still a problem. Lots of people are being bitten by this in the last 24 hours.

> game design is a rabbit hole

Game engine design is a rabbit hole :)

Game design is the ultimate lockbox - you're unlocking an entire imaginary world which has some platonic existance in your mind.

And since you mentioned Luanti, it deserves to be much better known as a credible open alternative to Minecraft. You could do a lot worse then designing/prototyping your game with Luanti as the game engine.

https://www.luanti.org/

The styles of Cyberpunk 2077 and Red Dead Redemption 2 are also dead givaways about their training data. There might also be a whiff of the Witcher 4 demo in one sequence.

The interesting possibility is that all you may need for the setting of a future AAA game is just a small bit of the environment to nail down the art direction. Then you can dispense with the army of workers to place 3D models on the map in just the right arrangment to create a level. The AI model can extrapolate it all for you.

Clearly the days of fiddly level creation with a million inscrutable options and checkboxes in something like Unreal, or Unity, or Godot editors are numbered. You just say what you want and how you want to tweak it, and all those checkboxes and menus are disposable. As a bonus that's a huge barrier to entry torn down for amateur game makers.

Liquid Sky soundtrack:

https://www.youtube.com/watch?v=T5UxwohjhHw&list=PL9F0ACA601...

Synth genius. I actually have it on vinyl.

> [LLMs] spit out the most likely text to follow some other text based on probability.

Modern coding AI models are not just probability crunching transformers. They haven't been just that for some time. In current coding models the transformer bit is just one part of what is really an expert system. The complete package includes things like highly curated training data, specialized tokenizers, pre and post training regimens, guardrails, optimized system prompts etc, all tuned to coding. Put it all together and you get one shot performance on generating the type of code that was unthinkable even a year ago.

The point is that the entire expert system is getting better at a rapid pace and the probability bit is just one part of it. The complexity frontier for code generation keeps moving and there's still a lot of low hanging fruit to be had in pushing it forward.

> They're great for writing boilerplate that has been written a million times with different variations

That's >90% of all code in the wild. Probably more. We have three quarters of a century of code in our history so there is very little that's original anymore. Maybe original to the human coder fresh out of school, but the models have all this history to draw upon. So if the models produce the boilerplate reliably then human toil in writing if/then statements is at an end. Kind of like - barring the occasional mad genious [0] - the vast majority of coders don't write assembly to create a website anymore.

[0] https://asm32.info/index.cgi?page=content/0_MiniMagAsm/index...

SEEKING WORK | Remote | Onsite San Francisco Bay Area |

Versatile back end code and infrastructure development services. Typical deliverables are Dockerized microservices specified by OpenAPI, written in Go or Python, with a REST/JSON API and backed by PostgresDB or MySQL. CI/CD pipelines included and deplyment on any of the major cloud vendors. And of course your custom business logic.

If you have more complex needs, e.g. project or team management please reach out as I have extensive career experience in all aspects of of technology development and management in startups and large enterprises. Need a contract manager or CTO perhaps?

At minimum I can help you develop your back end infrastructure from the ground up. Basically I offer development of back-end componenets you can slot into your Kubernetes or Docker environment from day one.

On the business side, you get a fractional developer for a no-haggling fixed monthly rate, corp-to-corp billing, long term support for your code, careful vetting of dependencies for licensing and security, and a professional approach to your technical needs.

Contact: contact@nulladmin.com

Website: https://nulladmin.com/

LinkedIn: https://www.linkedin.com/in/abnulladmin/

You may want to look at this: A foundation model to predict and capture human cognition

https://www.nature.com/articles/s41586-025-09215-4

Maybe that should be the future of K8s 2.0. Instead of changing the core of the beast tweak it minimally to get rid of whatever limitations are annoying and instead allocate resources to put a hefty AI in front of it so that human toil is reduced.

At some point you won't need a fully dedicated ops team. I think a lot of people in this discussion are oblivious to where this is heading.

Let me throw this out as a provocative take:

In the age of AI arguments about a configuration language are moot. Nobody is going to hand craft those deployment configs anymore. The AIs can generate whatever weird syntax the underlying K8s machinery needs and do it more reliably than any human. If not today then probably in 3 months or something.

If you want to dream big for K8s 2.0 let AI parse human intent to generate the deployments and the cluster admin. Call it vibe-config or something. Describe what you want in plain language. A good model will think about the edge cases and ask you some questions to clarify, then generate a config for your review. Or apply it automatically if you're an edgelord ops operator.

Let's face it, modern code generation is already heading mostly in that direction. You're interacting with an AI to create whatever application you have in your imagination. You still have to guide it to not make stupid choices, but you're not going to hand craft every function. Tell the AI "I want 4 pods with anti-affinity, rolling deployment, connected to the existing Postgres pod and autoscaling based on CPU usage. Think hard about any edge cases I missed." and get on with your life. That's where we're heading.

I'm emotionally prepared for the roasting this comment will elicit so go for it. I genuinely want to hear pro and con arguments on this.

Reznor released 5 movie scores last year. He's got 2 Oscars, 1 Emmy, 4 Grammys, and 3 Golden Globes the latest one this year. Multiple nominations in all those categories. He's far from done.

Apologies for the hijack. RIP Brian.