HNHacker News
TopNewBestAskShowJobs

claytongulick

1,306 karma · joined April 14, 2018

submissionscomments
claytongulick··on Flet 1.0 – Build cross-platform apps in Python
Looks really interesting, but I notice that Bluetooth isn't in the supported services.

Sort of limits how useful it would be for me - Bluetooth is one of the only reasons I'd do a native app instead of PWA.

claytongulick··on Why I'm still bearish on LLMs after Navier-Stokes
And the 6 year old doesn't cost more than the GDP of a medium sized country.
claytongulick··on Blizzard Workers Win Historic Union Contract
I don't think "raw, unfettered capitalism" and unions are orthogonal.

Actually, quite the opposite - in a free market, labor is free to organize in any way they want. If they have enough leverage over the employer, they'll succeed.

The anti-free market part is the regulatory quagmire that the US has in place for organized labor.

I think that's the thing that most people take umbrage with.

I don't know for sure, but I've heard from knowledgeable folks that it is more effective in Europe.

claytongulick··on Cloud in a Bottle: making self-hosting accessible to everyone
Thanks for bringing this up, I was going to ask the same.

I haven't heard any complaints about NextCloud, they've added some enterprise features, but all in a positive way, from what I can tell, not detracting from home and personal use.

claytongulick··on VC isn't VC anymore
Is it different than the crypto days?

I was part of a couple raises back then and I remember every founder was being forced to find some blockchain application, no matter how inappropriate, to get any attention from the fad-tech VCs

claytongulick··on VC isn't VC anymore
Interesting, I'm curious about the changes - I know AI has drawn most of the pump & dump sharks, just like crypto before it, but there were always a set of responsible holdouts that resisted fad-tech for real value opportunities.

Are you seeing a difference there?

claytongulick··on VC isn't VC anymore
I have so many thoughts about this, I'm not sure where to start.

Doing a raise has always been weird, there are lots of things that impact it, and different players in PE have totally different theses and motivations.

As a founder, finding the right investment partner has always been one of the most important and difficult things.

As a rule of thumb, I recommend to founders that only about 50% of the value of the investor is their cash investment. In many cases, less than that.

The things that are at least as important are their advisors (who open doors for you), their portfolio companies that can partner with you, and the alignment of their thesis and worldview.

I don't see any of that as having really changed much recently, other than a tightnening of capital for non-AI companies, but I suspect we're going to see a big shift there over the next 18-24 months, as the pressure from the LPs to deploy stays the same, but fingers get burnt from this bubble.

Also, don't forget family offices and industry VCs (Optum Ventures, etc...) that have a lot of these features built in to their structure, not just the fund.

claytongulick··on No AI Fridays
> But when you sit down at an awesome synth and start jamming, do you care whether the firmware was written by a real live human?

Yes. Very much so.

I want to play an instrument that was thoughtfully designed and created at every level by someone with deep knowledge and passion.

I want hand crafted waveforms that someone spent endless hours with a headphone making just right, and they were damn proud of it.

I want to see videos of the person who did it and how. What inspired the sound, how it evolved.

I want the human experience.

claytongulick··on No AI Fridays
> Hear hear. I do not understand the people who bemoan the loss of slide rules and books of trig tables in the era of graphing calculators.

Why do you think so many people prefer to listen to live music over an album playing on a jukebox?

Live music doesn't have the polish and perfection of a studio recording, it's much less efficient to have to manually play instruments than to just hit "play" on a song. So why bother?

I think it's because we're looking at things wrong all over the place.

Similar to live music, I definitely prefer to read code written by a real human. I can get in their head, understand how they were reasoning to solve a problem, even if there were problems with their reasoning.

When I hear all the breathless takes about AI "productivity" improvements, my brain just circles aback to the old saw about activity over achievement.

What are your outputs? What is activity and what is achievement?

A business person brags that AI is able to generate a report for them that used to require lots of manual work linking spreadsheets. OK.

What was the value of the report? What was the goal? I like to follow this all the way down-

- Widget sales report

Why?

- To inform inventory and production decisions, as well as marketing

Why?

- So that the company can make money? So that the founders can be proud of their product? So that the product can improve lives in some way?

Why?

- So that the people working can have improved lives, and/or more status in society

Why?

- So that collectively by competing with each other in structured ways, humans as a whole thrive and benefit

Etc...

OK. Where does AI fit into this? What has it taken away from the human experience, and how had it affected all of these goals?

The goal isn't code. It isn't even the product that the code makes - it's the reward signaling that you get from cooperating and competing with other people.

When you short circuit your efforts at creation by using AI, it feels the same to me as saying you're putting on a live music show, but instead just hitting "play" on an album.

claytongulick··on Judge rules Trump administration’s blacklisting of Anthropic was illegal
Congress ceding it's authority to the executive branch is an issue of national concern, has been going on for decades, and is definitely not limited to one specific party [1], though I do agree that the current administration has been testing the limits of executive power more than any other that I can recall in my lifetime.

[1] https://www.brennancenter.org/our-work/analysis-opinion/its-...

claytongulick··on EPA says power for data centers can sidestep pollution laws
The law was specifically written to exclude things like independent power sources and backup generators etc...

From 42 U.S.C. § 7651a(17)(A) [1] which defines the power generation units that are covered by the law:

"(17) (A)The term “utility unit” means— (i)a unit that serves a generator in any State that produces electricity for sale, or (ii)a unit that, during 1985, served a generator in any State that produced electricity for sale."

The key phrasing there is "for sale". The clarification from the EPA simply restated that these units don't fall under the regulatory authority because under a plain reading of the law, they are not generating power for sale. [2]

The EPA certainly cites Trump's power agenda several times, but that doesn't make the plain text interpretation of the law wrong.

[1] https://www.law.cornell.edu/uscode/text/42/7411

[2] https://www.epa.gov/newsreleases/epa-issues-permitting-guida...

claytongulick··on EPA says power for data centers can sidestep pollution laws
Are you sure you're aiming your ire in the right direction in this case?

The EPA provided guidance that the CAA (Clean Air Act) itself was written to not apply in these cases.

Shouldn't you be angry at the authors of the law?

claytongulick··on Kagi added a setting for removing paywalled links from search results
I remember at the time they were working on it a lot of people (honestly, including me) were pretty critical of them "wasting" their limited resources on AI stuff. We saw it as being sort of similar to the t-shirt drama [1].

I'm here to admit I was wrong, it's incredibly useful. It misses the mark frequently, like any LLM, but it's well annotated with footnotes that link to the original material.

I use it all the time.

I also love how easy it is to use or not - just add a question mark to the end of your query to use the AI, or leave it off for traditional search.

With that UX it doesn't feel like it's being shoved down my throat.

[1] https://blog.kagi.com/celebrating-20k

claytongulick··on Gemini 3.7 Flash
This happened to me once when I was setting up some virtual servers for a startup.

I wanted a typical dev/qa/prod with medium specced boxes.

I was denied for quota, with an esoteric process for review.

I'd just made a case for deploying to GCP over AWD so got a bit of egg on my face. Went over and had it done on AWS in a few minutes.

A couple days later, the Google product team contacted me. I told them what happened.

It got escalated, and I ended up on a call with like 5 or 6 people from Google, some very senior. I told them what happened.

They made very concerned sounding noises and told me how this was a product failure on their part, how they'd get it corrected, etc... and they'd fixed my account so I could now make the machines. Of course, I was already deployed to AWS at that point.

That company grew and ended up with a pretty big cloud spend eventually. Google totally missed it.

I was at a new startup a few years later and decided to deploy to GCP.

Denied for quota.

claytongulick··on Chartreuse, a case study in how scarcity and authenticity can drive desirability
For the curious, the Chartreuse liquor isn't the original formula, it was developed in the late 1800's, IIRC.

The original "secret elixir of eternal youth" from the crusades era is still available though, it's meant to be consumed as a medicinal tonic with a spoon. It comes in a cool small wooden bottle, I just scored and got three of them at my local Total Wine [3].

Edit: I just realized most probably aren't familiar with the history here. It's a fascinating story that starts with the crusades, involves the monks feeling both France and Spain, smuggling the recipe through prison, etc... [1] [2]

[1] https://www.chartreuse.fr/en/story/

[2] https://en.wikipedia.org/wiki/Chartreuse_(liqueur)

[3] https://www.totalwine.com/spirits/liqueurscordialsschnapps/h...

claytongulick··on Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows
Are you also paying 8,000 employees [1] and funding massive infrastructure [2]?

[1] https://www.makerstations.io/openai-employee-statistics/

[2] https://www.wheresyoured.at/oai_docs/

claytongulick··on Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows
> The major AI labs are gross profitable when selling access to inference.

Do you have a good source for this?

claytongulick··on How to Survive in a Louisiana Swamp
I'd really love to read an article that actually talks about how to survive in a Louisiana swamp.

This article is written in the style that I find particularly tedious - long winded overly descriptive character portraits.

More of a novelette than an article.

I don't think it's an attention span thing for me, it's more that as I try to read it, all I can think of is how the author is purposefully adding to the word count to drive those ad impressions.

I have to try to sort through endless exposition and excruciating detail to try to learn something interesting, all the while being fed an endless diet of ads. And tracking cookies. And difficult tracking options. And pop-ups.

No wonder people just want to use AI to read and summarize things.

My AP English teacher would have flunked me for writing like this.

claytongulick··on I'll be stepping back from leading product for X
I recently had my account hacked - it was a pure lurker account, I only scrolled elephant videos and such, and never posted or did reactions or anything.

Someone logged in to my account from a state I'd never logged in from before and posted a bunch of spam for an hour.

X perma banned my account. They told me I was banned for life, and any account I ever tried to make in the future would be banned. I appealed, explained that my account had obviously been hacked. It was denied.

So, ok.

I went and canceled all of my business subscriptions to Super Grok, and removed any business relationship I have with them. I've been banned for life, after all.

After a couple months, I have to say, it's been a total positive development. I miss the elephant videos, but I absolutely don't miss the constant background anxiety from political outrage/engagement farming. I liked grok, got some laughs from unhinged mode, but don't really notice not having it anymore, lots of good options out there. Hell, I get 90% of what I want (better search experience, summaries) from Kagi.

It's wild to think that people are going to be relying on X for financial stuff now.

Hope their accounts don't get hacked.

claytongulick··on Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs
I think (hope?) we will.

Check out the New Luddite movement [1] [2]

[1] https://www.cnn.com/2025/10/08/business/ai-luddite-movement-...

[2] https://en.wikipedia.org/wiki/Neo-Luddism

claytongulick··on Eight Myths on Software Engineering and GenAI
Oh, some of us see a massive disruption coming, just maybe not in the way you think.

Talented, experienced devs leaving the field out of disgust would be pretty disruptive.

Crumbling, faulty technical infrastructure with unmanageable heaps of AI slop that no one understands would definitely disrupt businesses.

Skill atrophy, and an entire generation of "developers" that have never actually written code will be disruptive when something breaks and no one knows how to fix it (this isn't a hypothetical, I'm seeing this phenomenon in a lot of large enterprise groups).

The total elimination of novel solutions, new ideas and new approaches to development and the resulting industry wide stagnation won't be disruptive, but will certainly be a drag.

claytongulick··on Ten advances in mathematics and theoretical computer science
> And then they come up with another thing that needs to be solved in order to prove it is important/hard/impressive. And once that happens, they do it again. And again. That's what "moving the goalposts" means.

The fundamental argument that I've personally made since the early days of this is that LLMs are not reasoning, in the way that word is commonly understood.

There are lots of reasons why that argument needs to evolve that could certainly appear to be "moving the goalposts", but let's take an example.

A lot of AIs were tripped up by the question "Should I walk or drive 50m to the carwash?" Several folks liked to use that as an example that illustrates that LLMs aren't reasoning, but as the models have been trained on that specific example, it's of course less useful. An AI can mostly nail it now.

So a different example is needed. A new demonstration of how these things fail at basic reasoning a child can do.

Did I move the goalposts? I don't think so. The fundamental argument stays the same. It's not hard to find lots of examples that trip up LLMs, because they are what they are: statistical inference machines. Nothing more and nothing less.

Useful, sure. But also commonly misapplied to areas for which they are inappropriate solutions.

claytongulick··on Ten advances in mathematics and theoretical computer science
> People keep saying this. Why?

For the same reason that you can't draw a 15 of Diamonds from a regular card deck.

claytongulick··on Progressive Web Components
Try keeping them in the light dom.

Everything works like normal html + css, but you get all the benefits of reusability and encapsulated script logic.

Lit-html (the library, not the framework) gives this approach super rendering speed too.

claytongulick··on Launching Health in ChatGPT to US Users
A Business Associate (BA) under HIPAA is a entity that contracts with a Covered Entity (CE) using a Business Associate Agreement (BAA). OpenAI wouldn't be considered a BA unless they had an executed BAA with a CE.

There are a wide variety of things, including communication with providers, that OpenAI can do without being regulated by HIPAA. The provider, however would most likely be covered, so they would be under an obligation to not disclose any PHI.

As an example, consider a hypothetical "email my doc" app. This app could take health information from the patient and email it to all of their doctors without falling under HIPAA at all.

What you're probably thinking of is the HIPAA Privacy Rule (there's more to HIPAA than just that, the Security Rule is pretty involved). Under the Privacy Rule, CEs and BAs can not disclose PHI without a patient's consent, with a few exceptions.

If the patient is disclosing that information though, all bets are off (aside from state level regulations).

claytongulick··on Launching Health in ChatGPT to US Users
HIPAA only applies to data originating from a covered entity.

It doesn't apply to information the patient supplies directly.

As a patient, I can take out a street sign displaying my health records if I want. It's my data.

If I want to upload it to a strange AI or any other app, that's completely fine. HIPAA isn't involved, for either me or the company I upload to.

In fact, quite the opposite. CMS has been going to great pains over the past decade or so to create standards to help patients get access to their records and own them themselves.

Different states may have additional protections in this case, and I know there has been some discussion at the federal level about regulating these "wellness" apps.

As an interesting technical detail not many people know, HIPAA only applies to Covered Entities and Business Associates engaging in the electronic exchange of information using "standard transactions", and was intended to regulate insurance and billing transactions.

A doctor that keeps paper charts and bills cash directly to the patient wouldn't fall under HIPAA at all.

claytongulick··on Jelly UI: Soft-body physics for native HTML form controls
I love the aesthetic, and I really love that they implemented it as a web component library.
claytongulick··on Heresy (2022]
Losing your job, being "cancelled", publicly vilified and unhirable is not the same as being executed, so therefore we should dismiss it as a false equivalency?
claytongulick··on AI boosts research careers but narrow the span of ideas explored: study
LLMs are the tech in question, not ML (AI?) in general.

LLMs are fundamentally limited by their architecture to only return a token predicted by a statistical inference, essentially lossy decompression.

It's like arguing that taking an image, compressing it with JPEG and low quality, then decompressing it into something blurry with some random color values thrown in is creating new art.

No one is arguing that everything a human has created is good. No one is arguing that LLMs can't be useful.

Sutton is arguing that it can't be novel. Cherry picking a couple of words doesn't change his argument, which is very clear.

claytongulick··on AI boosts research careers but narrow the span of ideas explored: study
Alternatively, you can walk away from your career in disgust, taking your skills and experience with you, as many people are.

Should be interesting to see what happens to the programming profession when there isn't anyone around anymore who actually knows programming.

← PreviousPage 2 of 34Next →