Yes, I can read and comprehend English and you should assume I read the page. Because of the "At least" wording, I was curious what a person who has actually used the feature has noticed, aka, learning from people who have actually done it already.

Are you using AWS RDS Custom to receive the WAL Streams or are you using something like Pigsty? Really curious about the actual specifics

fantastic!

> the model download is orders of magnitude greater than downloading the browser itself, and something that needs to happen before you get your first token back

sure but does this mean the model is lazily downloaded? that is, if I used this and I am the first time the model was called, the user would be waiting until the model was downloaded at that point?

that sounds like a horrible user experience - maybe chrome reduces the confusion by showing a download dialog status or similar?

also, any idea what the on disk impact is?

If you didn't imply that, I apologize.

If you did mean that, I disagree with you precisely because your point works if you only care about dependency management - it falls apart on system state. A static binary is a process on the host and shares the same process space, network stack, and filesystem.

OTOH, a container is a jail (the primary usecase): I can't cgroup a static binary's memory usage or give it a virtual network interface without reimplementing "container lite". Containers aren't just 'statically linked programs' - they allow me to use the kernel as a hypervisor for isolated environments.

What they are though, a messy but practical compromise to Unikernels - which was my last point in our GP.

Personally commiting to using Tailscale as a core foundation of my infrastructure and Ionscale is my hedge against getting Hashicorped.

> Service discovery is basically just Docker's internal DNS. Caddy-docker-proxy can use it to find healthy upstreams

Do you have a writeup of this somewhere? I'm unaware of being able to manage Docker's internal DNS over some kind of an API (would appreciate if you know a way to). The only way I know is to manipulate network aliases via Docker Engine API. As a result I use Hickory DNS with RFC 2136. That coupled with Caddy-docker-proxy gets me extremely close.

What surprises and gotchas did you have to deal with using k3s as a Kubernetes implementation?

Did you use an LB? Which one? I'm assuming all your onprem nodes were just linux servers with very basic equipment (the fanciest networking equipment you used were 10GbE PCIe cards, nothing more special than that?)

If you had said "unikernels" I would have had no arguments to make.

I don't you made that argument but could a valid conclusion of your comment be that, because Kubernetes is so ubiquitous, using it frees you from being a Sysadmin/SRE?

No argument there. The Toyota 5S-FE non-interference engine is a near indestructible 4 cylinder engine that's well documented, popular and you can purchase parts for pennies. It has powered 10 models of Camrys and Lexus and battle proven. You can expect any mechanic who has been a professional mechanic for the last 3 years know exactly what to do when it starts acting up. 1 out of 4 cars on the road have this engine or a close clone of it.

It's not what any reasonable person would use for a weedwhacker, lawnmower, pool pump or an air compressor.

- using Tailscale SSH is brilliant

- using caddy-docker-proxy for ingress is brilliant

What do you use for:

- service discovery

- secret store (EDIT: Crap you use Infisical. No shade, I just have this horrible foreboding it will end up like Hashicorp. I use Conjur Secretless Broker but am tracking: https://news.ycombinator.com/item?id=47903690)

- backing up and restoring state like in a DB

PS: Have you been having issues with Hetzner the last few weeks?

Would it be realistic or reasonable to detect a header like `X-kloak-ENABLED` or specific endpoints in the case of HTTP?

Similar for wire protocols like PostgreSQL or gRPC?

Our would a usermode proxy be easier but not preferred due to overhead?

- What's the best way to discuss this specific topic with you? As an https://github.com/spinningfactory/kloak/issues or something else?

- My specific usecase is to not need Conjur Secretless Broker (https://github.com/cyberark/secretless-broker) - my understanding of eBPF is entirely superficial but from a 30k ft view, it looks like this can not only replace it but would be a far efficient solution (Conjur would be a user-space proxy while kloak would be at lower levels of abstraction)?

It's perfectly fine for you to say non-Kubernetes isn't either your focus or on your 90 day roadmap :)

- At the end of the day, a business is an activity that allows a person(s) helping another person(s). Even in a strictly amoral environment, voluntary trade is a positive-sum game where both A and B gain positive utility: A provides a positive utility to B by giving up something A had but B didnt (B has more than before) and B compensates A for it by giving up something B had but A didnt (A has more than before).

- Due to the profit incentive, there will be scammers, grifters and similar people who are not interested in the "helping" part but rather the "profit" part

- A more powerful person can dominate a weaker person

- Societies, of which organizations/companies are part have hierarchies of people

- A person "higher" up the hierarchy has control and influence over a person lower down that hierarchy

- A significant motivating factor for a person starting a business is independence (be at or towards the top of their local hierarchy) and ownership (be at the root of their local hierarchy)

- The person paying the money out of free will has control and influence over the person accepting the money out of free will in exchange for the help

- Regulation is the forced injection of a very large organization asserting themselves into the premier position of the hierarchy (For example: If regulation bans a specific chemical solvent, the supply chain must either shut down, find an alternative, or go "to the black market"/"go underground" where the regulatory authority cannot inject themselves)

- Regulation is hard to change and it's entirely a legal construct

- Regulation is rarely deprecated (For example: there are regulations in place that were written for an era where people would start their cars by hand although its been decades since anyone has actually done that)

- Normal people don't understand legal constructs any better than they can understand multivariate calculus or multithreaded, concurrent code

- Competition is great for customers. Anything that increase competition should be encouraged.

- We must treat customers as adults who are not challenged nor disabled. To keep our discussion manageable, we will ignore those scenarios where customers are not adults or otherwise challenged and disabled

- There are economies in scale but barring regulation or hostile takeovers, scale can only be achieved by willing, cooperating parties (For example: Say there are 3 phone companies in the US: A, T and V. If either one of them wanted to have majority stake in the "one" merged company, short of the other two conceding willingly, that "one" merged company will never exist)

- Inefficiencies can hide better at larger companies

- Larger companies are risk averse although surprisingly they are in the best position to absorb risks - aka Innovator's Dilemma (For example: Google invented transformers but management decided against it because they were afraid of hallucinations affecting search result quality. 3 engineers at OpenAI read the transformer paper and the rest is history. Google is still playing catch up!)

- A significant motivating factor for a person starting a business is independence and ownership that allows them permissionless innovation

- Innovation is an overall gain for humanity which is why society rewards innovation and penalizes rent seeking.

- People can use knowledge to do both evil and good

- Humanity optimizes for the good of people over time

If you disagree with any of the above, let me know which, ideally with supporting arguments.

Now back to our discussion:

> I say that 'Big HVAC' will in the future outcompete all the small shops; that it hasn't happened yet says nothing

It seems like you prefer rejecting evidence that doesn't support your hypothesis. My recommendation is you do the exact opposite. I'm engaging with you precisely because your take is absolutely counter to mine and I a know for a fact, because I have done this before, that my understanding will be much improved at the end of our conversation. It will provide me even more clarity than I have.

> Yes, regulation can heighten the barrier of entry (but also lower, although less common)

and why do you think it's less common for regulation to reduce the barrier of entry?

> You're naming companies in industries that have become even more consolidated than when those companies were the behemoths

So your argument is:

- Sears: Retail is now even more consolidated

- Blockbuster: Entertainment is now even more consolidated

- IBM: Software or Hardware is now even more consolidated

> They are behaving rationally and in their self-interest

You're conflating the rationality and self-interest of an entrepreneur with that of an employee.

Moral and company alignment aside, a researcher in LLMs might not care much whether they go work for Google or OpenAI or Anthropic. They might not even care whether they go work for Alibaba or Bytedance. OTOH, sama is definitely not going to work for Google or Anthropic and it's absolutely in his rational self-interest to behave that way.

There are even more inconsistencies in your writeup but I think this is a good checkpoint for me to allow you to reflect.

I also suggest we continue this over email (details in profile). Thank You for continuing to engage, I appreciate your POV.

The reason why Silicon Valley is the hotbed of such innovations, large booming businesses is regulators haven't yet got their hands into software.

In comparison, more regulated industries like BioTech, Education, Finance, Oil, Gas, Infrastructure is a joke.

The Biden administration had started the ball rolling on shackling up AI and LLMs: regulating, enforcing, mandating bottlenecks. It would have been kneecapping ourselves - essentially placing on ourselves the very sanctions we now place on China! The U.S. wouldn't need China to invade Taiwan to bring down our economy, we would willingly do it to ourselves! The push to regulate AI through mandatory reporting thresholds and compliance bottlenecks for large compute runs is the exact type of barrier-to-entry that kneecaps open-source disruption and entrenches the current giants. Hopefully these illinformed suicidal tendencies are on pause, for now.

I CANNOT for the life of me, understand why we would think, placing any limit on our creativity and productivity is a smart decision but I digress.

> do you think in a fully free market, without any governement intervention, those mom and pop shops would not be bought up or merge?

Is there a regulation right now that stops BigHVAC from acquiring your local, community 1 to 5 person HVAC company?

We don't have to engage in hypotheticals. I already asked you "Are there small businesses in your area like the 1 or 5 person HVAC or cleaning or landscaping or breakfast company that you disdain as much as you disdain UnitedHealthCare?" and I don't have a response yet.

> The truth is that consolidation is an attractor. If it hasn't happened yet in a market, it will. Economies of scale give an undisputed advantage

Mathematically, probabilistically, a software engineer will make way more risk adjusted money working as an employee at the MAG7 than starting their own startup, yet, every month, we have software engineers joining and creating startups.

A similar mindset makes your HVAC tech start their own or continue to run their own even though mathematically, probabilistically, economies of scale would give them an undisputed advantage - but that, consolidation isn't free. An even remotely skilled operator understands that and what that entails.

The moment a big, consolidated HVAC company raises prices or provides terrible service, a smart technician can buy a van, print some flyers, and steal their customers. Free markets allow the mom-and-pop shop to constantly respawn. Infact, Gavin Newsom has managed to distort the HVAC market in California by making it really onerous to operate refrigerants that does make it pretty hard for mom-and-pop shop to constantly respawn. This is why you will notice a lot of PE picking off HVAC companies in California!

By your logic, the world would have stopped at Sears, Blockbuster, and IBM. Yet, the first two aren't even a thing in 2026 and I don't want to comment on the IBM of 2026.

There are even more examples than these that obliterates the tired "consolidation is an inevitable endpoint" argument. A reasonable person would immediately acknowledge that size does not guarantee survival in a market where innovation is permitted and the entry fee is low.

The "feudalism" that you apparently abhor is actually caused by government intervention. When governments heavily regulate industries (like education, healthcare, finance, infrastructure), they make compliance so expensive that only massive, consolidated corporations can afford the lawyers and lobbyists to survive. Regulation protects the consolidated giants from the mom-and-pop startups!

I must deal with reality though - your sentences suggest companies always consolidate and the only way is to counter that with regulation. You really believe in this and this does hold true in heavily regulated industries but then you need to ask why does the same not hold true for industries with low barriers to entry and permissionless innovation?

Here's a thought experiment: Before the AI boom, the MAG7 rivaled the GDP of entire nations and dominated global market caps. They certainly don't need AI and LLMs to stay happy - infact Google, who invented transformers, is now rivaled by OpenAI and a swarm of startups founded by the very engineers who left Big Tech to disrupt it. This isn't the only time this has happened in tech!

So why are they now rushing to outcompete each other, over multi billions of dollars, to the extent they are self-sacrificing staff to freeup capital to invest into compute?

This certainly isn't true of small businesses and the local mom-and-pop shops in your area.

For free markets to work, small businesses need the ability to thrive and starting one should be exceptionally easy.

What has unfortunately happened in a lot of developed countries, is that in order to regulate the very large companies you're upset about, governments have made it nearly impossible to start small businesses to act as competition. That makes it even easier for companies to consolidate and evolve into a monopoly, strong-arming everyone - the exact opposite effect of what those regulations were hoping to fix.

Are there small businesses in your area like the 1 or 5 person HVAC or cleaning or landscaping or breakfast company that you disdain as much as you disdain UnitedHealthCare?

This might entirely be true but I'm hoping that's because the frontier models are just actually more expensive to run as well.

Said another way, I would hope, the price of GPT-5.5 falls significantly in a year when GPT-5.8 is out.

Someone else on this post commented:

> For API usage, GPT-5.5 is 2x the price of GPT-5.4, ~4x the price of GPT-5.1, and ~10x the price of Kimi-2.6.

Having used Kimi-2.6, it can go on for hours spewing nonsense. I personally am happy to pay 10x the price of something that doesn't help me, for something else that does, in even half the time.

I am from India and have friends who are immigrants from Russia, China and Cuba. We don't take lightly to being lectured about communism. We didn't move to the U.S., the bastion of capitalism, because communism had worked well for our grandfathers and parents and continues to do wonders for its society.

You're describing a standoff at best and a horrible parasitic relationship at worst.

In the worst case, the supplier starves the customer of any profit motive and the customer just stops and the supplier then has no business to run.

This has happened a few times in the past and is by 2026, well understood as a way to bankruptcy.

That has always been the beauty of free markets - it's self healing and calibrating. You don't need a big powerful overseer to ensure things are right.

Competing with customers is a way to lose business fast.

For example:

- AWS has everything they need to shit out products left, right and center. AWS can beat most of their partners and even customers who are wiring together all their various products tomorrow if they wanted. They don't because killing an entire vertical isn't of any benefit to them yet. Eventually they will when AWS is no longer growing and cannot build or scale any product no matter how hard they think or try. Competing with their customers is their very last option.

- OpenAI/Anthropic/Google isn't going to start competing against the large software body shops. Even if all that every employee at TCS does is hit Claude up, Anthropic isn't going to be the next TCS - it's competing with their customers.