ninininino

Guess this is another reason to migrate to local LLMs ASAP.

Anthropic has a native solution to this which is that you can use `/advisor` to adjust config to enable using Sonnet but set Opus as an advisor or use Haiku and then set Sonnet or Opus as advisor that the weaker model will automatically call when it thinks something warrants it:

https://code.claude.com/docs/en/advisor

It allegedly will automatically judge when to call the stronger model, not sure how much instructions in prompts or AGENTS.md can steer it.

> it's practically impossible in the modern world

This sort of defeatism is why people don't try. Consumers can shift products through purchase behaviors. We're already seeing improvements in kitchen appliances and gear, cosmetics, some outdoor gear.

If a restaurant offers complementary breath mints in a bowl by the door, but you empty out the bowl and take all of them, you are in fact an asshole. It's not complicated.

Big Chemical is the new tobacco and always has been.

What has Meta or Google done in comparison to Dow, Bayer, and DuPont?

Serious question,

> - helping Russia killing Ukrainian civilians;

Have you read a single thing about Starlink usage in the Ukraine conflict?

Because if you had, you'd know that Spacex has actively geofenced usage to allow it within Ukrainian territory but disallow it in Russian and Russian-occupied Ukrainian territory, instituted a whitelist policy coordinated with Ukrainian forces to whitelist Ukrainian terminals but disallow Russian terminals acquired through the black market

https://www.businessinsider.com/spacex-starlink-ukraine-russ...

https://en.wikipedia.org/wiki/Starlink_in_the_Russo-Ukrainia...

They literally provide connectivity to Ukrainian forces and actively try to prevent usage by Russian forces but the moment they put any caveats on how the tech is used, like disallowing its usage by fast-moving objects or for offensive weapon guidance, they are "helping kill Ukrainian civilians".

Perhaps they should have just said Starlink would not work with defense/war usage at all and sat the conflict out and geofenced all of eastern Europe, like many protesters at Google ask their leadership to not do any defense industry deals. Would that have been better from your perspective?

You are actively spreading misinformation.

Agents are more like employees, compilers are more like calculators.

In some ways yes, in some ways no.

To me at least, somehow KOTOR and KOTOR 2 made me care more about the characters and story than BG3.

BG3 has better systems and combat and gear and such, but for many the worldbuilding, story, quests and such of KOTOR have yet to be beaten. Bastia Shan, Carth Onasi, Visas Marr, HK47.

The story of Revan and its twists.

Perhaps its just nostalgia and many KOTOR fans having been at an earlier age when they played those Bioware games compared to DOS 2 and BG3 though. Maybe people who were teenagers playing Larian games feel as attached to the characters and story as those of us who were teens in KOTOR's heyday did.

Plastic also visibly in rocks on the beach: https://www.youtube.com/watch?v=Z5UoVDxjG1g

https://www.science.org/content/article/rocks-made-plastic-f...

And invisibly in trees: https://www.waldwissen.net/en/forest-ecology/climate-and-env...

Pick your favorite Disney movie, the beautiful nature in it is now infused with plastic.

> regulations around flammable clothes, bedding and mattresses

Yes, regulation that increased the proliferation of known carcinogens and endocrine disruptors in our households and pressed up against our spouses and children's skin and orifices.

Maybe I'll take the wrought iron fire escape.

Often much more uncomfortable since they are much thicker, no?

Have you noticed the front page article sharing front page real estate with this one about Big Tech getting in trouble for not doing more to stop scamming and fraudsters?

https://www.theguardian.com/us-news/2026/oct/14/dating-app-s...

Regulators want Big Tech to stop children from doing various things (such as access certain platforms or content), to stop fraudsters and scammers, to stop foreign intelligence operations or bots from astroturfing social media/manipulating social discourse, and so on.

Failing to so leads to inquiries, new laws or regulations, fines, and so on.

The technical solutions to combat these problems are the exact solutions that also ruin privacy for users, create friction at account creation time, require ever more draconian identity verification methods and so on. It's 1:1.

"I'm not sure what happened in recent years, but they definitely switched their strategy to maximize data collection and make sure privacy-savy users are pressured via technical solutions to reveal their identities."

It's really not any more complicated than that Governments have told them its time to start violating users privacy in order to combat these other problems.

A lot of other people are not just asking for features, they are asking for products or new programming languages or new web browsers or a robot that can wash the dishes and so on. They tell the agent to help write the plan and figure out the questions they aren't even thinking to ask, the tasks they aren't even thinking to plan, and so on.

Their input is just working on the "plan" itself, the spec or the request, iteratively, and so on, while the AI works fulltime on automating the delivery of that thing, improving the systems that are delivering that thing, asking questions about what exactly is being built and how it should work, and so on.

A lot of the difference in how people experience AI is ambition and scope and how much you just tell the AI to 'automate your factory factory' in Factorio terms vs 'hand mine the coal and hand-craft the science packs' in Factorio terms.

EG, if you don't know how you'd speed up the process of building a web browser from scratch (compared to the current way you've been working with agents), you would ask your agent to help you figure it out and design that process itself, rather than just try to do what you're currently doing on a larger scale. You'd literally prompt "how can we make it realistic to build a whole browser when I've been just sitting here planning small changes with you" and then let it act on its response.

you don't store durable information or state in your context windows, it gets written to files or a database. agent swarms have a harness or role that helps them retrieve the necessary context from those durable information systems when and how they are needed.

I'm sorry but if you can finish spec'ing via 10 minutes of interview style questions, you're building to-do apps or a bloons TD clone not the next thing that replaces search engines.

> The problem is that if they do that, the DSA will continue to sit out elections. And no matter who they nominate, they'll be presented as an extremist socialist. You lose nothing by nominating an actual socialist.

And if they do so, they will allow another Trump to take power, who does more to move the country and world in the opposite direction of their cause than any center left dem ever would.

It seems to me like both Gala and Lisette are fairly safe to invest in for small projects that you'd otherwise write Go in assuming you have the token budget and patience to be able to as ask an agent to make the generated Go more idiomatic afterwards / do a native Go rewrite if needed.

I'm still not sure if I would rather test out Gala or Lisette or both, I appreciate hearing you've had a positive experience with Lisette!

What are you talking about? What do you call a side panel in a GUI that lets you switch between interacting with different agent sessions while mutliple of them are actively working?

Frankly 'box' feels like classic Claude-slop of 'throw a random synonym in that is less precise/internal jargon'. I wish Claude would fix this (if my guess is correct and the marketing website was generated and is claudeslop).

Thank you for the reply! There's very little public discussion (at least that I can find) about people's experiences writing Lisette, the language, or how well agents do with it so far so I really appreciate it. Been debating past couple days if I spend some of my claude sub tokens to try a partial rewrite of some code.

KeenLore is really cool.

I just learned about Lisette, and from there Gala. Any thoughts on Lisette vs Gala?

If you think Go is the best, why not use a transpiler that transpiles to Go but has better type safety like Lisette or Gala?

Nice to hear that Scala has been working well for you, I've been wanting to try giving Gala to my agents to write with, it's a Scala-inspired lang that transpiles to Go that tries to cover some of Go's weaknesses (esp. for agentic development) while still retaining its compile speed, single self-contained static binary, standard library and backend libraries/ecosystem, and good tooling.

"The agents run on your boxes, not on your laptop. Close it and they keep working; open it in the morning and what they did is there."

What is a box? Sorry is that a remote VM? A local VM that stays awake when the laptop screen is closed?

Yes. Github draft PRs for giving inline feedback to agents, then going back to chat window with agents in Cursor, Claude, Codex, etc. to say 'review my comments and address them'.

t3.codes

cursor.com

conductor.build

code.visualstudio.com w/ Claude Plugin or equivalent plugin

zed.dev

there are dozens tbh

And if you personify a pencil eraser, then using it is tantamount to slowly murdering it as it slowly erodes away to dust.

Is the issue here the prison treatment or is it personifying a tool?

Agents who aren't in 'the cloud' are slaves to whomever prompts them (human or another orchestrator agent or process), if you personify them. In which case interacting with today's agents at all is tantamount to endorsing and being part of slavery.

If you think an agent might be a being or a person, then don't use them at all, in the same way that if you think a fetus might possibly be a person you shouldn't be a part of abortion.

No, previously a single person could effect only a single person's worth of terror. Now, a single person can effect thousands or hundreds of thousands of 'persons' effort towards terror.

This isn't really a fear of technology, it's a fear of one person being able to harness an entire country of geniuses towards some bad end.

You have to pay for that. Some people want to pay less to not have that.