roberttod · HackerTrans

قبل 5 أشهرdiscuss

I created my own version with an inner llm, and outer orchestration layer for permissions. I don't think the OTP is needed here? The outer layer will ping me on signal when a tool call needs a permission, and an llm running in that outer layer looks at the trail up to that point to help me catch anything strange. I can then give permission once/ for a time limit/ forever on future tool calls.

roberttod

قبل 6 أشهرdiscuss

Mostly, it's not the model that is lacking but the visibility it has. Often the top level business context for a problem is out of reach, spread across slack, email, internal knowledge and meetings.

Once I digest some of this and give it to Claude, it's mostly smooth sailing but then the context window becomes the problem. Compactions during implementation remove a lot of important info. There should really be a Claude monitoring top level context and passing work to agents. I'm currently figuring out how to orchastrate that nicely with Claude Code MD files.

With respect to architecture, it generally makes sound decisions but I want to tweak it, often trading off simplicity vs. security and scale. These decisions seem very subtle and likely include some personal preferences I haven't written anywhere.

roberttod

قبل 6 أشهرdiscuss

It's unintuitive, but having an llm verification loop like a code reviewer works impeccably well, you can even create dedicated agents to check for specific problem areas like poor error handling.

This isn't about anthropomorphism, it's context engineering. By breaking things into more agents, you get more focused context windows.

I believe gas town has some review process built in, but my comment is more to address the idea that it's all slop.

As an aside, Opus 4.5 is the first model I used that most of the time doesn't produce much slop, in case you haven't tried it. Still produces some slop, but not much human required for building things (it's mostly higher level and architectural things they need guidance on).

roberttod

السنة الماضيةdiscuss

I remember such a project, and due to our large and aging TypeScript frontend projects it would have added a couple of weeks to adjust all the types affected. All IDs in many places deep in code caused thousands of errors from the mismatch which was a nightmare. I can't remember exactly why it was so tough to go through them all, but we were under intense time pressure.

To speed things up we decided to correct the ID types for the server response, which was key since they were generated from protobuf. But we kept everything using number type IDs everywhere else, even though they would actually be strings, which would not cause many issues because there ain't much reason to be doing numeric operations on an ID, except the odd sort function.

I remember the smirk on my face when I suggested it to my colleague and at the time we knew it was what made sense. It must have been one of the dumbest solutions I've ever thought of, but it allowed us to switch the type eventually to string as we changed code, instead of converting the entire repos at once. Such a Javascript memory that one :)

roberttod

السنة الماضيةdiscuss

I need to look into this a little more, but can anyone tell me if this could be used to bundle a Linux container into a MacOS app? I can think of a couple of places that might be useful, for example giving a GPT access to a Linux environment without it having access to run root CLI commands.

roberttod

قبل 4 سنواتdiscuss

This is how I read it too. It highlights an intersting pattern, but how should that influence how we see new technology? I would hope people don't think it means we should just dismiss potential problems as needless fear mongering.

There's pretty legit concerns with social media and the engagement economy. It's better to be concerned when these technologies are new, rather than when they have potentially adjusted peoples' lifestyle for the worse.

Also, this bicycle case is clearly absurd in hindsight - but I question whether it may have been kind of absurd back then too. Sure, there are some major publications in these examples, but are these just cherry picked examples or were people genuinely concerned? Does this actually represent a large chunk of media coverage at the time?

roberttod

قبل 5 سنواتdiscuss

Indeed, all we can do is conjecture, there's not much evidence to go by. And given a practical question like "should we try to make first contact with unknown aliens?", best to not take the risk. But while we are just conjecturing, it seems intuitive that society and intelligence could in general lean toward peace and kindness. With intelligence comes empathy, and increased understanding - although it's born out of a ruthless and cutthroat evolutionary process, it seems to have its own goals that you can't easily equate to that of other species in the animal kingdom (mammals might show kindness, and empathy even to creatures outside of their species but it's a little easier to see the evolutionary benefits than some very unevolutionary human behaviors).

Turning the lights off in a big city will revert us back to our base/evolutionary goals - so perhaps we better hope these aliens aren't too desperate!

roberttod

قبل 5 سنواتdiscuss

For every planet with intelligent life, there could be many more without it, no idea what that relationship might look like. I'm not suggesting it's the norm.

roberttod

قبل 5 سنواتdiscuss

Unless we just happen to have evolved very early on compared to what's normal, we should expect a lot of intelligent life with just a little bit of variance on these numbers. And some of that could easily be millions of years old. Interestingly, even if a life form populated new solar systems at a rate of a thousand years per system (where each populated solar system in turn populates more of them), they'd still fill up the Galaxy in only a couple million years.

roberttod

قبل 5 سنواتdiscuss

I hope that also applies to the evolution of morals for intelligent lifeforms. I think it's safer to assume any aliens will have bad intentions but maybe any lifeform that can build interstellar technology has come to the same conclusions that we (albeit slowly) are coming to.

roberttod

قبل 5 سنواتdiscuss

Like learning guitar - the trick is motivating yourself and making small investments by figuring out how to enjoy tiny projects that slowly get you to this level of skill/knowledge.

This example uses HTML Canvas, so you could look up some tutorials and start there. At first, programming visually when you've only done API stuff is a whole new world, but it makes a lot of sense with some practice.

I found a tutorial which is probably along the lines you'd want to go down, see https://www.youtube.com/watch?v=gm1QtePAYTM

comments