Serge Sountsov retweeted
Couldn't agree more! Let's apply the best talent and merit we collectively have available to the technical problems we all share. There's no need or gain in constantly rubbing our political disagreements against each other either 🤘❤️
I would like to walk back some things, in the course of a short polite discussion, @dhh and I agree we both have flaws in our character. I have never had someone who I've personally made a target reach out to me for polite civil discourse. I explained to him in essence that just because you don't support a thing, doesn't mean you ostracize a large portion of the community. Support, agree or not, a large portion of the trans community, and other marginalized groups of people are extremely talented. Open source is for everyone. Much like the Internet of old where nobody knew who you were, unless you told them. As long as politics isn't brought into the project from left or right, real talent can make real change. This is what OPEN in open source means. I thank him for his humility and ability to have a polite and civil discourse, well acknowledging his own flaws, as well as mine. It may just be a winning option for everyone, that I cease dropping exploits, and actually start to code. If not for his reaching out, I could not have accomplished what I set out to do. Tribalism and Open source do not mix. Thank you all for reading.
Serge Sountsov retweeted
I've spent a good chunk of the past two weeks working on improving Netclaw's user-facing UX and it's agent-facing... agent... experience?
Building a "pit of success" for agents to find the right tools, right files, at the right time.
Some lessons learned:
1. Better to have a larger number of more focused, specific tools - don't create combinations of parameters that can result in illegal calls. i.e. have a "file_write" tool that can do both line-number based edits OR search and replace. Pick one contract.
1b. If you're worried about tool count bloating context windows, use dynamic tool loading. Netclaw already did this, so it made this transition easier.
2. Return strongly typed errors to the client with well-documented recourse and rescue. i.e. if the result of web_fetch produced a document larger than what we allow in the context window, send the agent a path to the scratch file where the output was saved instead and provide it with precise instructions on how to open it and how large it is instead of showing the agent a truncated result
3. Protecting the context window vs. introducing complex tool calling patterns is a natural source of tension and trade-offs. The risk is making it too easy for the agent to blow out its context window by consuming too much information too fast (i.e. bulk loading source files in the main agent window) before it's determined what information is pertinent and useful.
You want to set up guard rails to force the agent to either delegate bulk consumption to a specialized sub-agent or have it more gracefully search for the information it needs. You also want to avoid having the agent spin fruitlessly on long tool call chains too - this is what good evals + agent observability are for.
4. AGENTS.md guidance can help, but it's literally a suggestion at best.
Should still do it, but the meaningful enforcement of harness rules is going to come from the deterministic code around the tooling and policy layers. i.e. your shell tool should probably reject calls that don't set a CWD, the CWD-setting layer should reject or prompt the user for CWDs outside the permitted workspace, etc. And if you do reject an agent's tool call, see point 2 - offer a clear explanation + suggested recourse to immediately fix it in the tool calling output to the agent.
Serge Sountsov retweeted
They say: "protecting children from social media"
They mean: "anonymous users will be presumed underage and banned from speaking freely"
Serge Sountsov retweeted
The essential engineering cheatsheet of 2026:
agent → while loop
subagent → nested while loop
agent harness → the rest of the code
cloud agent → all the above, on EC2
Serge Sountsov retweeted
In Mere Christianity, CS Lewis has an awesome opening riff about how most people know the difference between right and wrong, but they justify acting immorally by appealing to "special exception." They know they shouldn't hit a friend, but what if that friend was being so mean? They know they shouldn't steal a seat a bus, but what if that person got up and created a moment's confusion and then the seat was up for grabs? Etc.
When I read this section, I thought a lot about contemporary politics and the way that people justify their politics, not by appealing to higher principles, but rather by appealing to "special exception" to argue that their admitted indecency is justifiable in context.
A lot of MAGA vice is justified by special exception. Trump's defenders rarely defend his crookedness directly. They don't say "it's wonderful to use trade policy to enrich the Oval Office, it's really awesome." They say: Well, look, it doesn't really matter, because the left is so dangerous, Biden maybe did something similar 3 years ago, Democrats would do the same in power, and so forth.
I heard something similar in that NYT conversation everybody's talking about. You even see it in the headline: ‘The Rich Don’t Play by the Rules. So Why Should I?’ Why, hello, special exception. When you start arguing that stealing food and French paintings is justifiable in the context of political protest in an age of prevailing distrust, you're similarly not arguing *for* any kind of a universal principle. Nobody actually wants 300 million people stealing fruit from the grocery store. Nobody actually wants every Louvre visitor trying to rip a Manet off the walls. These virtues don't scale. (Because they're not virtuous!)
Sap that I am, I want us to get to a place where politics is about fighting for what is right and decent, not about justifying what sort of indecent behavior might be somewhat understandable or technically justifiable given the other side's vice or the prevailing levels of indecency. The point is to build the kind of goodness that scales.
nytimes.com/2026/04/22/opini…
Serge Sountsov retweeted
I've looked into this more and this story is entirely 100% fake. A lot of people are as usual misrepresenting what happened here and not bothering to look at the details at all.
First, as I argue in my water post, there seems to be no physical way for Amazon to be responsible for even 1% of the increase in nitrates in groundwater that are contributing to cancer. I'll post that as screenshots in the following comment and link it. What happened here is big ag let a ton of nitrates leak into water to the point that they're polluting local groundwater and increasing the rate of cancer. Amazon's role here is they operate a data center in the area that evaporates a small fraction of the water and leave the nitrates behind, so they very slightly increase the concentration, but it seems impossible for that increase to be anywhere near even a few acres of a normal farm in the area.
But Amazon agreeing to pay this must mean they're guilty right? Not at all, the actual court case makes it look like the opposite is true.
In the settlement documents, Amazon states explicitly that it is entering the agreement "solely to avoid the burdens and expense of litigation," and denies "each and every one of the allegations of wrongful conduct. Amazon's public statement says their data centers "don't add nitrates to that water, and the water we return represents a very small fraction of the region's overall system." If you do the math, it seems impossible not to agree with Amazon's statement here.
Amazon is one of 17 defendants in this case. The others include the large farms and food processing companies (the groups that actually added the nitrates in the first place), and the wastewater facility that's been failing to deal with leakage into the groundwater. They've been building up the nitrate concentration in the water since the early 90s, way before the data center was opened in 2011.
Amazon was the first to settle because its marginal contribution to the problem is the easiest to pay off and move on from! The agricultural defendants are still litigating because their part in this is way larger.
But as usual people are entirely misreading this and now think that 1) A data center poisoned people and gave them cancer, and 2) It's so clear this happened that Amazon won't even fight it in court.
The info environment around this topic is completely off the wall insane.
Amazon agrees to pay $20.5 million to settle allegations that its data centers in Morrow County Oregon contributed to contaminating ground water with nitrate pollution in the agricultural community
Here’s what happens
- Amazon uses huge volumes of water to cool its data centers
- Because the data centers use water that has nitrates in it from the area, water evaporates during the cooling process
- What's left is highly concentrated nitrate water
- That highly concentrated nitrate water then goes back out over the farmland, seeping back into the groundwater water
Now imagine this nationwide on a massive scale due to thousands of data centers being built
Readers added context they thought people might want to know
This is being framed as entirely the fault of Amazon and their Datacenter, even though this was an issue long before AI, as reported back in May 2022.
oregoncapitalchronicle.com/2022/05/19/wat…
Serge Sountsov retweeted
Elizabeth Warren turned out to be one of the most profound disappointments imaginable as a Senator. Presented as a technocrat interested in what works, actually the absolute worst of the slopulists.
Warren says the housing bill does *not* have a drafting error, as Schatz said today
Rare Dem leadership split
“The policy is to block private equity from taking over the single family home, and that is quite deliberate. There are some folks in private equity who don't like that, but it's a very deliberate choice that is supported on a bipartisan basis by 90 senators.”
Serge Sountsov retweeted
Ruriko: I have spent *eight hours* attempting to sign into Microsoft on [the new] Chromebook. This is the never ending loop of hell.
Me: Oh let me try that. I will need your face.
Ruriko: My face.
Me: Yes see you’re heuristically falling into Extremely Hinky Behavior.
Serge Sountsov retweeted
I’m amused by American medical offices offering contracts of adhesion at check in, where staff gives a one sentence description of why they need a signature, directs you to electronic signature pad, and then your signature is affixed to a contract that you’ve never actually seen.
Serge Sountsov retweeted
An open letter to U.S. customers....
tl;dr:
(0) the tariffs are indeed that bad
(1) if you want a keyboard, buy it now; taxes on individual packages don't kick in until May
(2) why we wish we could onshore, and why we can't
(3) what you can do
shop.keyboard.io/blogs/news/…
Serge Sountsov retweeted
Just 81 signatures on this change dot org petition to get Reacher on Severance next season. Sickening how people just don't seem to care about progress
Serge Sountsov retweeted
KubeFleet is a cloud native multicluster solution designed to ease the deployment, management and scaling of applications across multiple Kubernetes clusters. Now a part of the CNCF!
thenewstack.io/kubefleet-the…
Serge Sountsov retweeted
Me: Ah a nice relaxing Monday where I can finally get some work done.
DOGE: Have you ever heard of checks?!
Me: %{*]% it.
Serge Sountsov retweeted
"How many American parents would feel comfortable letting their 9-year old take the L in Chicago or the subway in Manhattan? I don't know any. And as a result, you just don't see any unattended children do this. But in Copenhagen it's common place." world.hey.com/dhh/waiting-on…
