rsalus

I agree with your main observations, but the Costco example is a bit contrived. I expect companies to eventually figure out suitable applications for AI, and I doubt a flat subscription per seat will be one of them. Personally, I feel the main issue is that the tooling + systems needed to deploy AI successfully have only recently started to mature.

the frustration is largely because the overall substance is quite poor since it is typically imprecise by nature.

surely the supply of unified memory will rise to meet demand before this is needed

now consumption is becoming dominated by asset owners

absolutely true. I am not convinced that consumption can be wholly fueled by asset owners though.

the argument is productivity gains are increasingly driven by technological advances, which are spurred by capital investment. for example, if a company purchases software that increases their accountants productivity by 5x, should those accountants immediately be paid 4-5x more?

I would contend that the accountant should not - it should flow to who bore the cost of the input (capital owners). however, if you starve labor of those gains, it destroys the consumer base that capital relies on to buy its goods and services. therefore, society requires broad wealth distribution to function, which implies some level of redistribution by the state is needed.

yes, but wage growth has not tracked productivity: https://fred.stlouisfed.org/graph/?g=1X020

household expenses have been increasing without commensurate wage growth, resulting in lower savings: https://fred.stlouisfed.org/series/PSAVERT

2% is average. 1-1.5% is considered a slump, while anything over 2.5% is considered a boom. for instance, the post-ww2 boom (1947-1972) averaged 2.9%. at that rate of growth, a country's total output per worker doubles in roughly ~25 years.

amount of labor being generated per person has not really changed

not true, labor productivity has been steadily increasing: https://fred.stlouisfed.org/series/OPHNFB

workers are simply capturing less of the economic value generated by their labor.

no, he's almost certainly referring to labor productivity. https://fred.stlouisfed.org/series/OPHNFB

he's referring to labor productivity, e.g., the economic value produced per unit of labor input. it also _has_ changed significantly, as seen here: https://fred.stlouisfed.org/series/OPHNFB

however, unit labor costs has also been increasing (although they remain variable): https://www.bls.gov/opub/ted/2026/productivity-up-0-3-percen...

agree, from my POV the constraints are still there but we've optimized now. still haven't solved the core problems.

eh, they're a publishing business. choicescript is part of their business model - there's nothing inherently wrong with that.

so cool! how does this compare to ChoiceScript?

https://www.choiceofgames.com/make-your-own-games/choicescri...

they are forecasting $25-30B this year, so if anything they look to be neutral or even profitable.

The TAM for robots is much, much larger than for coding or services

how do you figure?

AI usage will directly impact said margins. Moreover, for the scenario you describe, companies need to have the capability to precisely estimate the cost of a given deliverable - not something possible with current tooling + models. You're also underestimating the market trend towards vertical integration: companies are not going to be constrained by a sector or niche. They will expand to capture as much value as they can, because now their capacity to do so is partially decoupled from labor.

It will certainly be a cutthroat landscape for engineers, but companies will be building _more_ capacity, not less. In other words, the demand won't disappear for skilled technical labor, it will just move higher up the value chain.

I don't know, even if AI allows two engineers to do the work of six, companies will likely just use that efficiency to expand their scope. I think we'll see short-term layoffs and a more stratified engineering field during the transition, but the fundamental need for deep technical expertise isn't going away.

tests are an important signal of course, but the use case you describe doesn't necessarily mean you need to follow TDD. the data suggests that creating the tests after the code is just as or even more effective, and at significantly cheaper input cost.

I've been finding enforcing integrations and behavior structurally (e.g., through codegen/schemagen, e2e tests, etc) more reliable than simply instructing the models to write tests. oftentimes these tests are pretty low quality anyway, and results in its own form of tech debt.

this is not really backed by any empirical evidence. there are simply more efficient means of verifying outputs than TDD.

for LLMs, TDD amounts to little more than ceremony. there is a study on this exact topic: https://arxiv.org/pdf/2602.07900

not necessarily, TDD has little bearing on output quality

I was a big proponent of encoding TDD red-green-refactor methodology into my agent workflows until recently when I made the same realization after reading this study: https://arxiv.org/pdf/2602.07900

TLDR; it found test-writing volume only weakly correlates with success and that encoding test-writing principles did not move resolution rates but _did_ materially change cost. Encouraging tests cost +19.8% output tokens for 0% gain; discouraging them saved 33–49% input tokens for ≤2.6pp accuracy loss. Separately, imposing the TDD procedure specifically seems like it can backfire: it actually _increased_ regressions from 6.08% to 9.94%.

IMO, where tests clearly help is primarily as an "oracle" applied after generation. It gives the models a signal that enables them to verify and self-correct if necessary.

agreed but depends on then language. for instance, the .NET equivalent (MemoryCache) is pretty poor.

not really true, SQLite supports WAL mode which allows concurrent writes (technically write _attempts_, but these writes are exceptionally fast and are serialized to the file-system anyway, so functionally equivalent to concurrent writes for p50 use case).

also, use-case for massively concurrent writes is pretty narrow, and SQLite is not optimizing for that anyway.

functionally they operate as a marketplace for cloud providers. I feel like there is value there, especially as API costs rise and companies explore cost-saving/efficiency. IMO, this is a particularly attractive value prop in the SMB space, where it is common to interoperate between multiple SaaS/software stacks.

there is a difference between concurrency in a distributed environment and concurrency on a single machine across processes. SQLite is incredibly useful for the latter.

you seem like the inexperienced one to me..

mirrors my own experience creating a persistent event log. I started with JSON, then JSONL, etc until finally landing on SQLite.

they need to make 5t-10t back, but not necessarily through selling tokens. as we can see, the frontier labs are making vertically integrated products. their revenue is no longer strictly tied to inference.

often times the classical Greeks + Romans would cite their family lineage using works of Homer and other poets

https://github.com/lvlup-sw/exarchos

It's an SDLC workflow harness for agents. Instead of using skills to encode my typical workflows (e.g., create PRD, then create plan using TDD, then dispatch subagents, etc) I've built a concurrent event-sourced process manager to handle it.