politician

Ugh. Most of this thread can be resolved by realizing that Teriyaki Bomb is using the word "friction" as a synonym for "constraints" and not "toil".

My guy, that's actually how fires are stopped.

https://en.wikipedia.org/wiki/Firebreak

https://en.wikipedia.org/wiki/Controlled_burn

Well this is ridiculous. Literature or music?!

No, learning to code is still worthwhile because the AI cannot do useful abstraction well at all. If you don't know how to code, then you'll fail to build useful tools and useful reusable components that can (a) further accelerate your development speed, and (b) reduce token spend.

As others have commented, this is an obvious application of event sourcing. It's irritating to see the claim of "deterministic replay" in the abstract along with the caveat "we can't actually do deterministic replay, so we store all of the model's responses and reproject off of that". Sure, ok, whatever. You're doing session recording and calling it replay.

Dario quit OpenAI to hype the AI apocalypse for quick cash and attention. Then, he walked right into an obvious crisis with the Pentagon by continuing to try to play both sides of the AGI doom story that even his own AI would've pointed out. Then, after being labelled a supply chain risk, he starts a new roadshow with the newest most dangerous AI model that definitely cannot be released to the public and its safer little brother Fable. A move that gets both his premier models shut down globally once the same government that labelled them a supply chain risk learns that Fable isn't actually safe from jailbreaks. Just prior to his planned IPO.

Dario might not be a literal idiot, but he might strongly benefit from training a model to do strategic thinking for Anthropic.

Try to automate the adversarial PR review-rebuttal loop "for productivity", so the back-and-forth between the AIs can run over night.

I think this is why reasoning chains and reasoning chain verifiers are so important. We need to be able to see an argumentation, not just an answer. The paper below goes into this in more detail.

HeavySkill: Heavy Thinking as the Inner Skill in Agentic Harness

https://arxiv.org/abs/2605.02396

Hahaha, ok yeah I see where you're coming from.

It helps me to think about it like a different type of function call. We've got normal functions, async functions, there's a Go project that turns HTML templates into "templ" functions. JSX functions. LLMs are just a new `infer` function type.

A few years ago if I suggested that you should write a program to help you with time tracking, I imagine that might get a few responses with pointers to some existing open source projects. In a few years, someone might point to support for infer functions in nightly rust.

In other words, I think that we're dealing with really poor packaging right now and it's stressful, and that in the future this will all be normalized and integrated into our existing workflows.

Hey Joe, I think the solution to your problem is in your post. You said that when you were tracking your time it killed your idea, and that when you stopped tracking your time you became unfocused.

Try letting AI classify your idea into a time-tracking bucket for you, and to generate a beginning of day report describing how you spent your time yesterday.

If you write down your idea, then it'll be harder to forget it. You can let the AI figure out where to put it and fix it the next day if it's wrong.

If you look at where you spent your time yesterday each morning, then hopefully it'll help you figure out a better place to spend your time today.

You can easily set this up with any harness. Just copy and paste my comment and tell the AI to make some skills.

It's clear that you understand that releasing a schema with this much documentation makes the AI-assisted reverse engineering loop trivial. I'm interested in understanding why I should trust/use/pay for your hosted runtime versus rolling my own to run on cheap VPS?

This doesn't make sense as a rebuttal.

Article is OK with people moving uncontrolled in/out/around Europe. Article is NOT OK with data moving uncontrolled in/out/around Europe.

It's stupid. It's quintessential luxury belief thinking.

The article supposes that there exists people that fear their data being misused but don't fear their taxes being misused on welfare for a random population of benefit seekers.

The article would have been better without the self-own "problematic baggage" language.

Which is mightily funny because in the opening paragraph the article equates "anti-free movement" with "problematic baggage". It's a problem if people can't move freely in and out of Europe, but not data -- that's our red line!

"ethical" is not a word that carries the connotation of a universally agreed upon set of behaviors. Different peoples, groups, and cultures vary in what they consider acceptable behavior.

I wonder how many gallons of polluted wastewater are discharged per day by overseas refineries. Does anyone know where Tesla stacks up in the global list of lithium refiners?

_underscore_ for italics conflicts with most identifiers in most languages.

Markdown was created in an era before the web had easily used components for structural syntax highlighting (tree-sitter) and where reliance on regex-based approaches was more common.

I think we'll be able to quantify sentiment from the data, and I look forward to doing so. There's a few other datasets that I want to look at such as whether there is evidence of participation suppression via rate limiting on a per-profile basis.

This is great. I've soured on this site over the past few years due to the heavy partisanship that wasn't as present in the early days (eternal September), but there are still quite a few people whose opinions remain thought-provoking and insightful. I'm going to use this corpus to make a local self-hosted version of HN with the ability to a) show inline article summaries and b) follow those folks.

You’re not missing anything. I’ve worked with many developers that are clueless about error handling; who treat it as a mostly optional side quest. It’s not surprising that folks sees the explicit error handling in Go as a grotesque interruption of the happy path.

In your understanding of the constitution, is the executive branch subordinate to the other two branches?

The OP said that they kept what they liked and discarded the rest. I think that's a reasonable definition for signal; so, the signal-to-token ratio would be a simple ratio of (tokens committed)/(tokens purchased). You could argue that any tokens spent exploring options or refining things could be signal and I would agree, but that's harder to measure after the fact. We could give them a flat 10x multiplier to capture this part if you want.

"Most of the code gets discarded." If you don't mind sharing, what's your signal-to-token ratio?

This is a decent approach. My concern with TDD is that writing tests necessarily implies designing an API upon which those tests operate. Here, the agent is instructed to "not write code, write tests", and yet, in doing so it defines an API. This will cause the AI to hallucinate the API. Layering in yet more tests on top of this will cause that API to deform in strange ways that pass tests but that the adversary will not be able to cope with because it runs too late in the VSDD process.

I've seen this exact process play out in my own work. The AI generates code and tests that pass with high code coverage and honors invariants set by spec. I look at the code and find a rats nest / ball of mud that will cost 10x more tokens to enhance should I ever need to add a feature.

So, I think you're on to something, but I think the process might be discounting extensibility and resilience under change.

Bingo. You don’t have to read much into this if you remember how the DoD uses the word trust. In their world, a "trusted" system is one that has the power to break your security if it goes wrong. So when they say "unrestricted use," the likely meaning isn’t just fewer guardrails it’s that the vendor doesn’t get to monitor or audit how the system is being used. In other words, the government isn’t handing a private company visibility into sensitive operations.

No individual, whether a colonel or a CEO, has inherent authority over national security decisions. Authority flows through democratic institutions. A contractor can choose whether to participate, but national defense policy is determined by elected institutions, not private executives. If society believes AI should or should not be used for certain military purposes, the venue for that decision is democratic governance not unilateral corporate refusal or approval.

On a CBS interview this morning, Dario defended his position with the claim that he must act because "Congress is slow." CEOs can and should make decisions about what their companies build or refuse to build. What they cannot do is substitute their judgment for the constitutional processes that govern national security. We must not vest de facto policy control in unelected corporate leaders.

The posters question was itself a deflection - and your response is moral blackmail. Why don't you answer my question? Why are you deflecting? See how that works?

Unlike Europe, the USA is not a technocracy. If you want to learn more about the 3 branches of government and the scope of their powers, ask your AI.

Are you an American? Do you understand that your safe easy life depends on a mostly autonomous nuclear deterrence capability maintained by the military you oppose? Deeply think about why you still have right to free speech, and what it takes to sustain those rights.

"This has never happened before." US could compel Anthropic to act; simply not doing business with them is restraint, not escalation.

This is ridiculous. These aren't unreasonable demands and the government has tools to compel tech companies to support the country regardless of any "collective action" shenanigans -- ask your AI to tell you about the Defense Production Act and the history of its use.

I simply do not understand why Americans tech companies and their employees will hew and cry about supporting the military. For those of you who support their position, have you ever stopped to consider that your safe, comfortable lives of free speech and protests and TikTok and food and gas and Amazon Next-Day deliveries is enabled by a massive nuclear deterrent operated by the very military you oppose?

It is just so disappointing to come here and read these naive takes. Yes, Anthropic should be compelled to support the military using the DPA if necessary.