thevinter

I struggle to see any value in this when DeepSeek is still a thing.

Well, to be clear, there was some feedback (e.g. it was calling subagents and printing some thinking messages).

That said, yes, it is pretty common for me to wait 10, 20 and sometimes even 30 minutes without steering the model or looking at what its doing. I usually write a pretty detailed prompt at the start that describes the issue, the usecases, the testing to do and the definition of success; then I sometimes ask it to draft a plan and give it a read, but after that I just press enter, let it run and come back when it's done.

In the end I do a manual review, both of the code and the functionality, but most of the time I get exactly what I wanted.

Personal anecdote: I exhausted my Claude usage yesterday so I decided to spend 20$ to try Kimi while I was at it. Logged in, paid, downloaded Kimi Code, set it to use K3 and prompted something along the lines of: "Check this repository and find all the settings that can be passed as input related to hardware, I/O, thread control or networking. Produce a report".

It thought for about 12 minutes and then told me I had exhausted my daily quota. (The next day Fable did the same task in 3m)

If you want to buy a plan for K3 do NOT buy the 20$ one.

Update: The Rust version has been released as Bun canary - running bun upgrade --canary will install this release.)

My biggest issue is that it’s impossible to engage with and give feedback on an AI written document

Thanks for giving shape to a general annoyance, in my case with code, that I only recently started noticing.

When discussing code with my colleagues, especially with those that I do not know particularly well, I often relied on the quality of the code they produced to modulate how technical the conversation ought to be be.

Now, instead, I often see very complex code from people that I know wouldn't have been able to produce it themselves, and I have no clue how to engage with it, how to review it with them, how detailed can my comments be etc.

It's pretty annoying.

One thing that I found interesting is that most of the discourse surrounding the topic happened with the assumption that the rewrite was happening with an Opus-like model, and not with Fable. Those assumptions, at least partially, were used as arguments against the fact that the rewrite was feasible and/or a good idea.

Clearly the model itself doesn't completely change the narrative, but at least as a note to myself, I would like to be more careful with assuming the capabilities of the models used internally by Anthropic and affiliated orgs.

And what happens once the "solid baselines" become unavailable for a reason or the other?

I appreciate (and encourage) using real life problems as a pretext for researching new topics and developing new skills, but

1. I'm really curious as to what's the desired outcome here? A spambot to flood people's notifications?

2. I'll admit that I have absolutely no idea how Instagram's API layer (and protection) works but wouldn't capturing the HTTP calls a more appropriate and easier approach to take?

Has rsync never broken features before, without AI?

It's not silly to have issues with something.

I absolutely understand and agree. As I said, I understand the underlying reason.

The silly part is the brigading - issues should be adressed on their own merits. The specific GH issue, and some of the comments therein, make the whole crowd they're affiliated with look bad. (imho)

This whole brigading is bizzarre and some people are behaving like irrational animals. I potentially understand the motivations that might bring one to want to "win" this battle but this really isn't it - it just makes you sound like a fanatic.

It takes 5 minutes to search for "regression" on the issue page and go through the 17 results. There are potentially even more on the tracker used prior to github.

I think this behavior is very silly and people are just trying to justify their hate to AI by latching onto every possible thing, seemingly forgetting that before AI people did mistakes as well.

If you have proof that AI involvement in rsync has lead to a significant increase in open issues please show it to me - I'll be happy to change my mind.

If that were to bethe case, a public scandal would be the last thing they need, and I would expect them to do everything in their power to keep it under wraps instead (eg privately settling)

Oh yeah sorry, I misunderstood the suit the comment was referring to.

It is my understanding that BAM took direct ownership of the local store and therefore the small claims case was also directed against them, but at the moment I can't find where I've heard that so I'm not 100% sure.

The lawsuit was *from* the previous owners against the new owners that kicked them out of the store.

Here's a video from the previous owners explaining their story: https://www.youtube.com/watch?v=zedmOopRTm0

No but I can link timestamps to relevant clips:

- At 3:06 they explicitly acknowledge the consignment and state they will be taking it over

- At 13:15 the CEO says he never had the LEGOs in the store and then is confronted with screenshots of said LEGOs from their official Facebook pages

- At 23:05 the new owner that took over the store (and also the LEGOs) first says he doesn't know about any LEGOs, then he says that he wasn't the one to sign the consignment and therefore doesn't have to give them back

- At 47:42 the same guy confirms again they have the LEGOs, tries to argue about the definition of theft and says that he won't give them back. (quoting "who cares if it's theft or not")

- At 49:46 the same guy admits again that he has the LEGOs and he promises to give them back if the actual owner provides him an apology and removes the negative reviews.

- At 1:00:45 corporate says "I'm not gonna distribute those things at this point. We've kept them on hold for this long so"

I feel bad for the guy who lost some LEGO sets. I do not like the podcasters and bloggers milking him for content for their media channels

The statements made by the company are simply untrue. And the guy who lost the LEGO sets (worth 100k$ btw) is directly working with the "bloggers" because they're his last avenue. He's also incredibly grateful to them because thanks to them he at least ended up winning in small claims court.

There seems to be a lot of misinformation in the comments, I would assume because the linked article doesn't cover many of the developments.

The youtuber Reckless Ben has recently covered the story and spearheaded a campaign of "provocative journalism" against the store[0]. Regardless of whether you support the way in which he goes about things, his video explains the story in much greater detail, and enormously expands on the malpractice of Bricks and Minifigs and the local police department.

Here are some bulletpoints in case you do not care to watch Part 1 + Part 2:

- Bricks and Minifigs explicitly threatened both the previous owners of the store and the original owner of the collection with lengthy legal battles

- The owner of the collection tried going the legal route but was quoted prices that he couldn't afford, so youtube was his last resort

- Bricks and Minifigs CEO publicly admitted of having the collection, being aware of the issue, and not wanting to give it back, while at the same time trying to run PR campaigns denying the allegations.

- BAM leadership went out of its way to create legal trouble for Reckless Ben, involving the police and fabricating false evidence about him

- The local police went out of its way to legally stop Ben, arrest him without probable cause, try to plant Heroin on his car, and even *ended up swatting his house*, dislocating his shoulder.

- All of this while the police department illegally scrubbed any incriminating evidence from the bodycam recordings they were obligated to provide.

This is an *insane* story that doesn't get enough credit. It not only exposes the inefficacy of (parts of) the American justice system, but also the enormous level of corruption and abuse of power of the American police (and tangentially the Mormon community)

I really recommend watching both videos. I promise you it's even more insane than it sounds like.

[0] https://www.youtube.com/watch?v=wscQpkcwgNU

All of these quotes are false and directly contradicted by publicly available statements made by the CEO + Corporate.

No, the guy tried to resolve the legal dispute with lawyers and has been quoted multiple hundreds of thousands of dollars in fees.

By what definition? While I would say that arguing by definition is pretty pointless, here's one from Cambridge Dictionary:

art (noun) - the making of objects, images, music, etc. that are beautiful or that express feelings

Gatekeeping the definition by excluding certain tools from being used in the creative process feels very silly to me.

There is definitely room for improvement: https://gist.github.com/simonw/88eecc65698a725d8a9c1c918478a...

Especially when it comes to detailed outputs or non-standard prompts.

I do believe it will get even better - not sure it will happen within a year but I wouldn't be incredibly surprised if it did.

Every time a new image gen comes out I keep saying that it won't get better just to be surprised again and again. Some of the examples are incredible (and incredibly scary. I feel like this is truly the point where understanding if something is AI becomes impossible)

It's a very simplistic and radical point of view that doesn't take into account the reality of the world we live in. It also doesn't take into account the intricacies of foreign politics and seems to assume that the gulf states are the only bad actors here. Finally "gulf states" is a catch-all so big that it's borderline funny. (What did Bahrain do?)

And no-one is preventing you from caring about those things. I build UIs with Claude a lot and I still spend a lot of the time thinking about the user experience and working with Claude to make an app as intuitive and easy to use as possible.

Probably because the person wasn't interested in planning their vacation and wanted just to enjoy the end result?

Let's not assume different people find the same parts of the process enjoyable.

I lived for months with a 4GB roaming plan. Given, I was not using it at home since I had a wifi connection, but I rarely came close to using all my data unless I was watching YT videos when traveling or something.

I share your sentiment and I agree we should be more mindful of people with metered/slow connections, but the last statement feels blown out of proportion.

Yes but isn't it a bit weird to be implying your customers are dogs?

Cursor came out 3 years ago. "Agentic" refactors have been a thing for 1.5 years. Vibecoding as a term has been created 1 year ago.

There are multiple companies that deploy to production daily. What are we even talking about?

Pi was probably the best ad for Claude Code I ever saw.

After my max sub expired I decided to try Kimi on a more open harness, and it ended up being one of the worst (and eye opening experiences) I had with the agentic world so far.

It was completely alienating and so much 'not for me', that afterwards I went back and immediately renewed my claude sub.

https://www.thevinter.com/blog/bad-vibes-from-pi

Are you intentionally keeping the benchmarks private?