deno

You made the model cry :(

The point of this was supposed to be "model tool use competency" not baseline over-eagerness.

Unless I'm missing something I think to a large degree you're just comparing system prompts.

If I add "Research the question extensively" to your prompt at the end I get the correct answer from Haiku and Sonnet Med on first try and I've reproduced the original prompt not returning the answer.

Unfortunately every other run now gets your gist in results.

Yes, you need 3f+1, so 4 for f=1, 7 for f=2.

https://lamport.azurewebsites.net/pubs/byz.pdf

But what if the third bookkeeper is malicious? You need at least four bookkeepers to achieve byzantine fault tolerance with f=1.

This would be hallucination rate and neither of those models excels in this area and in fact Grok does, or at least prior version.

This isn't tool usage competency, it's tool quality and/or luck. Regular web search is not good for grounding if you want accurate results. You can ask the model to make a tool for getting bus schedules and then use it only then you are comparing apples to apples in this case.

This likely comes down to how it accessed the bus schedules (i.e. web search tool) and not intelligence.

You need to add the actual bus schedule to context somehow (research agent, custom tool or just dump in prompt) and even the simpler modern models will be able to do the planning.

Don't assume it's only depression, check your testosterone levels etc., probably best to just get an endocrinologist consultation.

If you're tired and don't have energy you'll also feel depressed but that doesn't automatically mean you feel tired solely because of depression.

Imagine you created a torrent (and/or magnet link) with a file and then stopped seeding after some time. If it was popular it will probably live on, if not then eventually it disappears.

Don’t bring your sensitive data to airports.

I don’t want to get into specifics because ultimately it just comes down to the logical argument that just because a strategy is not by itself successful against all possible threats it doesn’t mean it’s worthless. By this logic you wouldn’t lock your front doors.

This is not relevant to memory encryption, after all the police could just plant any false evidence. You use video camera/CCTV and other evidence gathering to document such illegal police action.

Suing after the fact is a valid strategy and in free countries this would allow you to exclude illegally obtained evidence or evidence lacking proper chain of custody.

law in question: https://www.irishstatutebook.ie/eli/2017/act/11/section/7/en...

and recent Supreme Court decision that upheld its constitutionality:

https://www.algoodbody.com/insights-publications/password-pr...

Unless you live in North Korea, China, Russia, UK, France, Australia or Ireland it’s still illegal to coerce or force someone to give up their personal keys or passwords, so this feature is still useful against some law-bound adversaries in free countries.

You acknowledge your test is fundamentally flawed but then still use it for the basis of your conclusion. It’s worse not better that you were aware of this.

If the email is DKIM signed it could be easily proven

It still takes the same mindset and skills to use AI productively and effectively as regular programming. The productivity boost only applies if you know what you’re doing and can actually steer the agent carefully.

Vibecoding hits a glass ceiling very quickly and this will not be solved incrementally. Besides, if the agent could work autonomously to that degree then it would no longer need any prompting at all and we’re living in a very different world. On the other hand that would make the debt actually meaningless, so I guess that is 'a' solution.

I put the actual clause under, but let's forget the actual legal definition for a moment.

GPL license in spirit is about assuring the user freedoms. No user freedoms are limited in this case. You are free to modify and redistribute the software as you like. OrcaSlicer pulls changes from Bambu without any issues.

I don't think trying to enforce the license in this way, even if possible (which again I think if it was it would happen with Linux drivers long before), is the right thing to do anyway. All it's doing is painting the GPL as a liability to any business for no benefit.

AGPLv3 still has the termination clause which is at least in the worst case (total failure to comply) entirely self-contained.

I'm not however convinced they are really in violation by calling a binary plugin. GPL itself does not forbid you from dynamically linking to or calling unrelated software. The network plugin is analogous to a device driver, it's not core part of the slicer.

GPL differentiates between a "Combined Work" and an "Aggregate":

A compilation of a covered work with other separate and independent works, which are not by their nature extensions of the covered work, and which are not combined with it such as to form a larger program, in or on a volume of a storage or distribution medium, is called an "aggregate" if the compilation and its resulting copyright are not used to limit the access or legal rights of the compilation's users beyond what the individual works permit. Inclusion of a covered work in an aggregate does not cause this License to apply to the other parts of the aggregate.

If they tried to add DRM to Bambu Studio and prevent you from replacing their blackbox with a different one then that would be where they would clearly go against the v3's TiVo provisions.

You can pay the customs fees and local taxes yourself if you're shipping with Fedex/UPS, it's just a checkbox if you have an actual account with them.

Also you can certainly ship batteries via UPU network (Australian Post) but there are strict packaging rules you need to follow[1].

You don't really go into any detail as to why they didn't accept your package. Did you attach properly filled out CN23 label?

[1] https://auspost.com.au/content/dam/auspost_corp/media/docume...

SEV-SNP would actually be a very good outcome. But the Intel equivalent does not exist on desktop SKUs. Meanwhile there's a campaign to make sure every "P"C has a TPM module.

This will be used as reason to introduce remote attestation to games.

Building out for 100% of theoretical capacity makes no sense but you can still easily accommodate the small handful of power users with plenty to spare. Most ISPs will not drop or throttle users trying to get their money's worth if it’s fiber or similar. LTE of course that’s another thing.

That sort of horrible abuse only happens in areas where some provider has strict monopoly, but that’s an aberration and with Starlink’s availability there’s an upper bound nowadays.

Since I know how many of those businesses are run I'll let you in on the very obvious secret: there’s zero chance they have enough uplink to accommodate everyone using 100% of their bandwidth at the same time, and probably much less than that.

Residential network access is oversold as everything else.

The only difference with storage is there’s a theoretical maximum on how much a single person can use.

But you could just as well limit backup upload speed for similar effect. Having something about fair use in ToS is really not that different.

Exactly: https://news.ycombinator.com/newsguidelines.html

just click [flag]

Right, that's why I recommend linear switches. But they're marketed as a gaming switch and they don't deserve this harmful reputation. They're simply the only non-stupid type of switch.

Low profile scissors are a compromise. They are tactile but it's for once functional as it compensates for the obvious lack of travel. The result is a mediocre but still above average experience. You can type fast on them but with more fatigue than high-profile linear.

Marketers want you to believe there's at least three different distinct switch types for different purposes because they want to sell you the same keyboard twice or thrice. But in reality there's linear and stupid. And some of them make unnecessary noise. Some of them make really sweet nostalgia noises like Alps switches but it's still a worse typing experience.

This is the marketing mythos I was talking about. The best typist keyboard is regular linear switch. Typing without bottom out is impossible. Clicking is just audio. Of course most mechanisms of producing clicking mean some degree of tactility (added resistance), but any tactility bump you have to overcome means you come out on the other side with more force which means harder bottom out. The way to reduce bottom out force is to not have any resistance, which is what linear switch is.

The most popular type of switch is brown because it's essentially a linear switch that is not marketed towards gamers. It's just sad.

Most people and companies just use the keyboard that shipped with the computer. I don't think noise is as much of an issue as people make it out be.

Marketing made up this story about linear switches being for gamers. So now every mechanical keyboard needs to make unnecessary noise and offer extra resistance for harder bottom out or you're not a serious typist.

But that's not inherent to the keyboard. Linear switches are not any louder than cheapo office high-profile membrane.