threecheese

You aren't wrong, but this AI thing which we're trying to give people:

1. is something that many intensely don't want, and many with valid reasons; 2. will inarguably consume some of a resource which we all need, which is already becoming more scarce globally; 3. is in many cases contributing negatively to actual or perceived quality of life via DC noise and pollution; 4. threatens the livelihood of millions in the near term, even if it will improve outcomes over time; 5. is inextricably tied to and will substantially benefit a wealthy cohort, who is already seen as exerting undue control over populations governments ; 6. present potential systemic societal risk, yet is moving forward at breakneck speed; 7. has apparent total US government support, despite complaints about all the above, in a political environment which is frankly frightening, and where that government is exercising little restraint against - and collecting more sensitive data than ever about - citizens who protest against policy.

None of us should be surprised about this reaction.

Are you this guy? https://github.com/hofstadter-io/hof/tree/_next Third time I've seen CUE mentioned in this context in as many days.

SpaceX-stans (god i can't believe i just typed that) use it, there's a whole subculture grokking (heinlein is turning in his grave) their way towards the same vibe coded nonsense as the rest of us.

Tbh I think Thariq or one of the others passively hinted/suggested but didn't come out and make a categorical statement. I've seen anecdotal evidence of folks getting those extra data messages in various "not using the official harness" scenarios, but there's way more people like you who are not having issues.

If you say you wrote it, I'll read it. I commit to this.

We can all throw questions like this at our LLMs, and the objections you'll see here revolve around having our time wasted by folks who believe their LLM output is an accomplishment; we're getting sick of it.

Specs-as-history is rapidly becoming an antipattern as teams actually try to scale SDD. We're already putting guardrails in agents.md (non-normative in all caps), some teams are actually offboarding them - to Rally (which I find the height of hilarity).

Spec docs - at least in all the other things like this thing - aren't a living representation of truth, instead they are a collection of snapshots in time of feature deliverables. Layer 500 of these on top of each other and they quickly become not just useless but actively harmful. Archive them, compile them, whatever your plugin does, it still doesn't help. Nobody is going to pay for tokens just to keep the specs time-coherent, when the agent can just read the code.

Don't get me wrong, a spec is a useful bit of kit for development, just like Jira and Rally, but also like Jira and Rally they aren't the source of truth because there's already a better one which you just paid for.

One thing I've noticed is a lot more HF assets require personal information collection to download. Not sure if that's new corporate compliance or an attempt to manage bandwidth (or a marketing channel).

Why don't we just regulate software? Would accomplish the same end. And it's the same thing, at the end of the day.

Given this is obvious to mere mortals like us, how do economists defend the walmartization of every market? "It's working, don't fckit up", or "we're getting our bag, get yours"?

There isn't one adult who looks at this massive centralization and sees that it goes against the very principle of the thing?

My experience matches your claim, but I've seen others recently claim cli use (shelled out from another harness though) showing clear extra-data billing.

And I'm fairly sure Anthropic did implement that though, at least for Claude SDK (which is/was a wrapper around claude -p); I was using the sdk for a product, and had a demo to show/ensure it was billed via subscription. I basically unset any API env vars, ensured the cli credential was in place, and spawned agents and watched the logs. When I re-ran it during that time period, it clearly went to extra data (there are messages in response logs calling it out).

It supports json/jsonl output and verbose command/subcommand help; there's a lot of churn in "how do I create clis for AX" right now, I don't know if you can point to what's good or bad beyond "machine/digestible and progressively disclosed".

Recently did a quick experiment using four python cli libraries with codex, fairly big cli surface, well-typed code with docstrings, and the differences were negligible- it came down to those two qualities specifically, and the means to those ends didn't really matter. Obviously anecdata though.

As much as I would love to make the most of these tools, I am not confident that yielding control of my personal data won't hurt me - and this seems like table-stakes, based on the type of value everyone is getting from them.

I dont even know what should be private anymore - if US federal agencies can just buy ad data, join it with Flock, sprinkle Instagram in there, and inner join Social Security dump from doge, do I just give up and put the sex toys in the parlor window? Am I one OpenAI-sued-by-a-newspaper away from giving up the ghost?

"Attackers compromised 3 unique developer github credentials" is a little scary; this seems systemic - if three developers had their keys stolen, it implies the real problem is elsewhere.

They revoked all but one developers, which is probably the right approach, but I'd pin LiteLLM to "something before this version" and keep it there until further notice!

- A woman using Claude for her diary just had her information used against her. - A nytimes civil suit against OpenAI just had user data handed over to lawyers. - US federal agencies are buying ad marketplace data and joining it with other commercial and government data to extra-judicially identify persons

We are giving an unprecedented amount of personal data to these tools at the same time that the technology is being leveraged against people at an unprecedented scope and scale. No US legal framework exists to protect you against this. You should take this risk seriously.

The problem here is we're going to let the machines dictate the ontology, and we will lose expressiveness by reverting to the mean. All in order to communicate more cheaply with the machines - but on their terms.

Forgive the anthropomorphizing, it was necessary - "terms" has double meaning there, and each of us can understand that without an ontology - because we can access nuance. I can read schema, structure, and shape, and understand; but only one of those triggered some readers.

Anyway, the machines will choose terms which are most well-represented in the training data. This will normalize communication by chopping off the top-end where nuance lies. This is token-efficient.

Ontologies, in their implementation in the Linked Data ecosystem, do offer the ability to relate one understanding-system to another, but this is not token-efficient. I would love to see each domain, each business even, define their view of the world in the most expressive and culture matching owl representation, but I don't see it happening. "Shape" it is.

Dig into LinkML; I've been eye-deep in defining an ontology for my life (which I struggle to grasp), and just getting into the tools forced me to learn ... outward. There's so much cool stuff, and now I have a mental model as well as a machine-based thing I can talk to (obv an LLM) having a "shared understanding ".

Cheers. Same. (well, not with mrs frenck- that'd be scandalous!)

This makes the most sense. Meta can guarantee at least that cohort of sites will permit their agents to operate, and all they needed to do was publish a tos update. Some Meta PM is smiling today, in between bites of infant flesh.

I am so confused. You need to pay a tax to ... watch content of any kind? Like a book tax?

Or, is this "people watching broadcast content on devices", or "you'll get YouTube TV for free"?

Pets are OK sometimes! They do need care and feeding, though; some of us don't see that as chores, instead delightful hobbies. I can definitely see the perspective of "there are better things to do", especially if you don't like those things.

Once you really know your keyboard, it's like an extension of your hands. You can fly through those six terminals, and moving your hand to grab a mouse means stopping whatever you were doing. That wall of text is all familiar to me - I see an UI full of empty space and wonder what I could have used it for.

Of course there are great reasons for using pointers also. I'd say this is mostly a preference, I'm sure you are extremely effective using your own methods. I have maybe two dozen terminals open atm, and I can tell you what each of them is for - but put a phone bill or a bank statement in front of me and it might as well be Greek.

Not a reply to TT, but I just saw your nick and thought I'd respond. We both commented on a thread recently, and your response was really insightful and just a pleasure to read; I didn't take the opportunity to respond there (locked). You should, like, write a blog, man. I infrequently wish HN had a "follow" feature, but this one of those times. Peace.

ref: https://news.ycombinator.com/item?id=49768575

1. Entrenched in-company interests create friction by just existing - Ops/Risk/Support/Arch/ProcessEng. Staffing is on the knifes edge, no buffer. And so going fast $here makes you slow $there. 2. Customers expect defect-free software, and compromising quality == churn == revenue loss. 3. Token costs are already astronomical, and everyone who could already cut headcount. 4. Mythos scared the shit out of everyone, and every last cent is allocated to Enterprise Risk (another entrenched friction).

It’s a lose/lose proposition for every software company with more than three management layers. We're frozen in place.

I think there's a type of company for whom this works; a small team really can deliver a massive volume, but it cannot be constrained by your typical software bigco superstructure. Ops/Readiness/Process Engineering/Enterprise Arch/Risk etc, entrenched concerns with soft power through pure delivery friction.

You need to be an an org who is able to deliver directly to your consumer, who needs to be defect-tolerant with the expectation of quick iteration. That customer will have access to advanced tools that its competition is unwilling to adopt. Many will fail.

There are many verticals for which this does not work at all - but this is unlikely to be the permanent state. The new model will work its way through the market, and companies will fall as fast-moving competitors find a willing customer segment they can target, make that segment out-compete the others using their access to tools with superior abilities but higher defect rates. The customer base will eventually adopt, having become accustomed to this, and so slowly expectations will change, and so will our field.

I for one welcome our slop overlords.

If we released software with the number of defects in the popular coding tools, we'd be out of business before the end of the year. But they persist- for some verticals speed matters over quality.

I can't believe OP here. Where are these software teams, are they in the room with him? My teams are more stressed than ever: leadership doesn't understand why they aren't 5x more productive - even if they were 1.1x more productive, the ops teams aren't scaling, the defect management isn't scaling, the guy yelling about their Jira ticket status updates is just getting louder, and they are being pressured to reduce token spend while pressured to release more software.

Some motherfucker today suggested we should use Haiku for all coding tasks, and this guy is a Sr Director of real humans.

Same. I've spent 20 years trying tool after tool and finding some part of it unfit, some stupid edge case for my broken brain. I have so much of my own crap now that I'm building new crap to manage the crap. If my professionally-sourced anxiety wasn't eating my stomach lining I'd be having a great time!

It's all being built to try and capture the value of creating the new, good software. Aka a cambrian explosion of developer tools - primarily used to create developer tools and twitter attention - so people use your developer tool to create developer tools. And wikis - thanks to Karpathy, we now measure Obsidian clones by the gross.

In my professional life however, we are NOT producing more software ($bigco). We are on the edge of it maybe. My experience is we've been on the razors edge - just enough staff to meet the business needs - and so every AI iteration sped up some process, which exposed some limit in a dependent process, which we tried to free up with a reorg, which cut capacity somewhere else, and the snake eats its tail. We've had enough iterations to get to a place where an increase in developer output probably won't cause the whole thing to fall over. I hope.

But anyway, I spend a lot of time exploring new dev tools, it's sort of a hobby of mine; I like to understand the state of "what's out there" and this has helped me professionally. This has recently become impossible do the sheer mass of llm productivity, while at the same time more fun than ever; the amount of crazy interesting and seemingly mature brand-new things (in the dev space) has astounded me. I found Swamp Club recently, like wtf where did this come from? It's like five dudes with LLMs.

In this analogy, a programmer who cannot understand their workload is like a rug QA tester with eyes but not feet; or a safety inspector who doesn't know that the red dye is corrosive to skin, or that corrosiveness is possible, or even that it exists as a concept. Or that the "red" is a dye, and what a dye does.

Currently, learning programming is the lever to unlock architecture understanding. We will need a few new layers of abstraction before this changes.

As a non-js developer, Bun compiles my ts code to local executables nicely. Allows me to experiment with the new diversity of ts frameworks and distribute the binaries (hobby scope).