sickcodebruh

This misses the point about complaints of AI slop. It’s not about “worse”, it’s about “distracting”. It’s about AI fingerprints overwhelming the consumer. It’s not too different from hiring a freelancer who does a blatant stylistic ripoff of a popular artist in an unflattering way: the consumer of the work sees the ripoff first, focuses on the ripoff, is unable to appreciate the message being delivered. AI slop is overwhelming and distracting to the consumer in that it communicates messages of laziness, disinterest, disregard for the underlying message. Whether this is true or not in every case, we’ve seen enough cases where we have to start by considering it. It’s the vehicle for scams and spam. It’s like sending an email that starts with “FIVE TRICKS TO HAVE BETTER SKIN” or any other meme spam line.

A human freelancer might do work that is technically worse. Worse can be distracting in ways that are, well… worse. But it’s different, it doesn’t tap into a collective cultural understanding of “AI slop.” It may communicate even more poorly, but it’s not the same problem.

My wife and I met on original okcupid. Not much to say other than the system worked! It’s sad to hear they become tinder-y but I guess it’s hard to compete with the immediacy and dopamine prospects of swiping.

When we consider:

* LLM usage is new for the world

* Models are evolving quickly with high worldwide competition

* Hardware is evolving despite RAM shortages

Is investing a huge sum of money in equipment for local inference a wise use of money? Or are M5 Ultra and equivalently priced local inference hardware future-proof enough to be worth it relative to how the market is evolving? Maybe it’s all a question of what you’d spend otherwise on serverless or dedicated GPU spend…

Composer 2.5 is phenomenal for so many tasks!

Clerky is great. I used it a few years ago and not only was the web product as advertised, they were quick to answer questions and give advice where they could for years after. I hope they keep doing well within Stripe.

> Four score and seven years ago our fathers brought forth on this continent, a new nation, conceived in Liberty, and dedicated to the proposition that all men are created equal.

> The founding moment was not a new nation, but a Liberty-conceived proposition: all men are equal.

Wasn’t that a different level of challenge? It was easy to call out the problems with it and recognize gaps to close. With Seedance 2.5, I’m not sure how I’d describe it other than “looking like AI video and all AI video is inherently off-putting”. People smarter than me are working on this but I’d be surprised if it ever actually became tolerable to look at it.

The sample video looks incredible and the ability to keep details consistent for so long is impressive but it _still_ looks screams of AI in every single shot. I can’t put my finger on why, something about the way the rooms are put together, the expressions of the faces, the movements… It’s very unsettling.

This is great! Did a few rounds, will come back.

A few months ago someone on HN shared Dragonsweeper and it completely hooked me and my wife to the point that we had to agree to break up with it. https://danielben.itch.io/dragonsweeper

It is clearly not effectively communicated if the tool used to write it stamps its signature in such a way that it distracts from the great work they’re doing.

Crazy fast but the the way all header text sizes change briefly flash between values every time you switch tabs would drive me nuts if I was responsible for this.

Varied. Most were exactly the same processes I remembered from interviewing years ago: some startups with no processes and unable to get signals, big public tech companies abusing leetcode, awkward screeners hoping for keywords, and everything in between for the most part. Biggest change is they all asked about AI and wanted to hear that I embrace modern tools and am hungry to find the best ways to work. I am positive that skepticism about AI in November prevented me from getting some interviews; demonstrating fluency (insofar as one can be fluent in these racing waters) and excitement for it in January was key to getting some (but not all) offers.

Around 12 calls if I remember correctly. With the exception of those who ghosted me, every company was very prompt, respectful of my time, and had a reasonable process that went from first to final interview pretty fast. I got a ton of outright rejections to my resume, which I sent around a lot. Getting the first call was the hardest part. YC job board ultimately led to finding my new role.

The company I went with had the best, warmest process that reflected how they like to work and was built to find people who had overlapping priorities. Thinking back, I think that most companies either deliberately or inadvertently wind up with processes that sync up with how they are organized internally and what kind of people they hope to hire.

Obsidian is my favorite new tool. I started a new job last week and decided to spin it up on day one to help me onboard and learn their system. It’s been wildly successful, the missing piece I needed to really turn Claude Code into the learning and documentation tool I dreamed about.

Started looking in November, four offers by end of January, all decent, last two competing offers were fantastic with great companies and I accepted one. Past few months and even now I’ve had more inbounds from recruiters than any time since Covid boom. Offer salaries aren’t as high as Covid boom days but there are a ton of startups that need people.

TIL! I had no idea this was a thing. https://support.apple.com/en-us/111772

It really is a great fit for this feature.

The docs remark “VMs share the resources allocated to the user” so I interpret as resources allocated to your account, VMs provisioned within those limits.

My approach to Claude Code is evolving.

I’m still unable to get Claude Code to contribute meaningful features directly my large web app at work. Specs will sometimes help it get close but it eventually veers off course and enters a feedback loop of bad decisions. Some of this might be attempting tasks it’s not suited well for, or perhaps my specs just aren’t precise enough, but I had enough failed attempts that I stopped trying to do anything that I’d describe as “challenging” or need too much domain knowledge.

A friend recommended I try it for less brainy backlog tasks, especially the kinds of things I can run casually in the background and not feel too invested in. This keeps failure from being too frustrating because there’s minimal effort and success becomes a pleasant surprise.

My first attempt with this was writing Playwright tests of the large web app in a new workspace within the monorepo. It was a huge success. I explained some user experiences the way I’d walk a person through them, pointed it at a path on my dev server, and told it the process I wanted it to follow: use Playwright MCP to load the page and discover the specifics of using the feature, document execution steps, write playwright tests based on what it learned from discovery, run the tests and debug errors with Playwright MCP. I instructed it to seek out the UI code within the project and add data-testid selectors as needed. I had it write this process to a master task.md, then make more task markdown files for each feature to be tested. It was very effective. Some of the features were somewhat complex, requiring two users with two browsers interacting in non-trivial ways. Not 100% accurate and more complex features needed more contextual and code corrections, but overall it probably saved days of frustrating work.

There’s a really funny duality to mistakes in recorded art that is vastly different when viewed as a fan and the creator.

As a music fan, I really love little mistakes in incredible albums. They’re humanizing, they show that the recording was made by people and it makes the highs feel so much higher.

As an artist, I loathe mistakes in my own work and I will spend a basically limitless amount of time fixing annoying performance quirks in software — I’m talking things that I can do but didn’t get quite right — so I can listen to it without distraction or regrets. I know that nobody will notice these except me and the type of listener who does catch them will either not mind or appreciate it the way I would. But when it’s my own work, it’s different. I’m sure it’s the same for filmmakers so I understand the impulse to fix it later.

They are given fake names and identities in the platform in a deliberate intent to mislead the audience, deprive the real author of credit, and hide the real source of the work the major record labels. “Fake artist” is a generous term.

This is neat. I added myself.

I'm hitting a bug where clicking New York just refreshes the page instead of loading the city's page. Los Angeles, San José, and... most cities seem to exhibit it. Chicago seems to work?

Picking a city was tough. It insisted on the more general New York instead of the more specific Long Island City.

This is the main thing keeping me from trying it. If you could clarify how you protect and use our code, it would be a huge help.

So is Apple the building that contains a Starbucks or the business known as Starbucks? It seems like it’s both — that is a problem. Is Starbucks charging its customers steep costs for entering the building? Where does the existing App Store with its large transaction fees and restrictions around billing fit into this analogy?

I remember finding that secret track based on that "0 is also a number" and being blown away. IIRC there's at least one other secret song in another track's pre-roll.

The separation of art and artist is a complex thing that differs from person to person. I think one universal aspect is that one's willingness to divest from art (or product, in this case) is relative to the ratio of disgust felt for the artist to desire for the product. Elon has reached a point of cultural over-saturation while Tesla's products move in the wrong direction on the ladder of cultural cache, the result of their (Tesla vehicles') own foibles and the industry as a whole just becoming more competitive. While plenty of people won't buy his cars, I'm sure they will or would still use a Tesla Supercharger if they could, because the value there is simply too compelling and the relationship is brief but powerful.

Elon in particular has become a liability thanks to his deliberate efforts to be the face and voice of Tesla. For me and clearly others, it's no longer possible to think of Tesla independently of him. His products need to be that must better to work against it and they're just not making the cut.

Other products -- Shell, Nestle -- are reevaluated for disgust-vs-need each time. Most of us have tiny, brief interactions with these companies. There's no meaningful relationship, so I'd have to be extremely furious with a brand to avoid their product. Right now I'll get gas from Shell but not Lukoil, for instance.

This is a great walkthrough!

I love CSS Grid. When you’ve got a design that really needs it, it feels like magic. But truthfully, I so often try to use it but wind up back with flexbox. Or there’s some quirk to a design that just doesn’t work out. Subgrid will make it much more usable (at least as I understand it) and I’m very excited for it to have wider support.

https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_grid_la...

Cassettes are coming back in a big way. They never totally died in the underground metal world, especially black metal, but they’re bigger right now that at any time in my 20-ish years of recording and releasing music. They have a few benefits: extremely cheap to make and ship. Easy to store in a small apartment, car, pocket. More durable than a CD. Imperfect and grimy in a way that feels appropriate for underground music that celebrates rawness. My new album was released at the end of September and it’s selling well across all formats but the only ones where we’re sold out: limited edition vinyl colors and — you guessed it — cassettes.

After years of fighting it, I’m finally getting a new tape player for Christmas. I’m a digital audio guy but so many albums that I want to listen to can only be found on YouTube or tape and I want to support the bands directly. I can get a modern unit complete with rechargeable battery and Bluetooth for $160. https://www.wearerewind.com/ Finally, I can use my AirPod Max headphones to listen to a $6 raw black metal cassette as the Dark Lord intended.

One of my favorite experiences from my daughter’s earliest years was the realization that she was able to think about, describe, and deliberately do things much earlier than I realized. More plainly: once I recognized she was doing something deliberately, I often went back and realized she had been doing that same thing for days or weeks prior. We encountered this a lot with words but also physical abilities, like figuring out how to make her BabyBjorn bouncer move. We had a policy of talking to her like she understood on the off-chance that she could and just couldn’t communicate it. She just turned 5 and continues to surprise us with the complexity of her inner world.

I know that it takes a lot of effort to setup these things but I wonder about the long term benefit to buyers. They’re buying into the Launch Leopard Architectural Philosophy, which they may or may not appreciate and grow the way they hope. Today’s unproved framework or boilerplate is tomorrow’s tech debt liability.

Did you consider adding guaranteed support or consulting hours to help buyers launch and grow? That would de-risk the initial buy and provide ongoing value. “Buy the starter, get X hours of time, I’ll be there to help you as a partner on your journey.” Then a contract (hourly, quarterly, yearly?) for continued engagement. Boilerplates are everywhere, relationships and guidance are not. I imagine that anyone in the market for this would see that as a major value add and differentiator.

I encountered versions of this days ago. Someone I knew for decades died suddenly, unexpectedly, too young. Trying to learn what happened, I searched for his name and found YouTube videos and AI-generated “news” articles that spit out variations of the same public information. It was unsettling but also pointless, nothing but noise in a situation that was already so upsetting and confusing.

I didn’t realize Compose could now target iOS! That’s exciting. I haven’t written Kotlin or used Jetpack Compose since the before the latter’s 1.0 release but I was very impressed at the time. Kotlin in particular stood out as an expressive, powerful, fun language. Compose had some rough edges but I picked it up, new to Android and Kotlin, and was productive very quickly. I’ll certainly consider Compose Multiplatform next time I have the need.