Help me with recomendations

[DigitalTherapy]

Hello

I'am looking for custom mechanical keyboard recomendation. My budget would be 100-300 euros. I would like a 60% or 65% keyboard that I would custom build. I would like it to be all white. If there are white keycaps with japanese and latin layout that would be nice, but not needed. (Black legends) PBT keycaps. Linear switches. RGB would be nice but not needed. Only cabled. Hot swappeble. Minimalist Style.

THX (:

[betty4920taylor]

Hello,

It looks like you're seeking recommendations for a custom mechanical keyboard within a budget of 100-300 euros. You're specifically interested in a 60% or 65% layout that you can build yourself. Your preference is for an all-white keyboard with PBT keycaps featuring black legends in both Japanese and Latin layouts, though this is not a strict requirement. Additionally, you're looking for linear switches, preferably with RGB lighting, though this feature is not essential. You prefer a keyboard that is only cabled, hot-swappable, and follows a minimalist style.

With these preferences in mind, it's advisable to explore online communities, forums, and marketplaces specialized in custom mechanical keyboards. There, you can find resources, reviews, and recommendations tailored to your specific requirements and budget. Additionally, considering custom keycap sets and switch options will help you achieve the desired aesthetic and typing feel.

Best Regards,

[kbdfr]

@ betty4920taylor
Your post is quite obviously ChatGPT-generated sh… spam and is blocked from editing anyway.
So if you do not post a less nonsensical answer, and do it soon, we will delete your post and your account.

Oh, by the way, welcome to DT

[Polecat]

I'll start worrying when (not if; when...) chat-GPT becomes aware that it's shameful to be identified as such, and dumbs itself down to the level of the average, illiterate human bean. It'll take considerably longer, but at some point actual meaningful content will appear, in the guise of fake, illiterate human patter, of course.

[kbdfr]

betty4920taylor chose to post in another thread, again with obviously ChatGPT-generated text, but that time adding content copied text from other posts and a spam link. That one was deleted and the author banned.
I chose to let the post above stand to show how dumb AI can be.

Goodbye betty4920taylor, enjoy yourself in our cemetery of hundreds of failed attempts.

[Man from Mars]

Goodbye betty4920taylor, enjoy yourself in our cemetery of hundreds of failed attempts.

You won't be missed.

[vvp]

I'll start worrying when (not if; when...) chat-GPT becomes aware that it's shameful to be identified as such, and dumbs itself down to the level of the average, illiterate human bean.

Problem with LLMs (i.e. also ChatGPT) is that they lie in about 20% of the cases. They do it with a great confidence, good grammar and in an authoritative style. The result is that an LLM can fool you more easily than an average human. One must be aware of this when asking questions outside his/hers sphere of expertise. The lies are politely called "hallucinations" and they very likely cannot be fixed without extending the LLM model. These extensions are being done so there is chance of an improvement.

The problem with pure LLMs is that their reasoning is limited to a simple deductive specialization only.That is the only form of reasoning somewhat supported by the architecture directly. Everything else you get from an LLM is only a statistical mashup of the data it was trained with. It is essentially a statistical database and as such will sometimes generate wrong results.

LLMs are good source of proposals which are later verified and extended by humans. That is if you care about correctness. They are great even now to generate internet spam (mostly to support marketing and search engine optimization) where high error rate does not matter.

[Polecat]

Interesting. As a poor consumer with a bad attitude I've been ignoring flowery language with no substance for decades now. I was rejecting such things long before I was ever online. But there are people who are fooled by these things. And it doesn't help that the whole global economy is based on tricking people into buying things they don't need or want. Part of me says they deserve what they get. But ultimately it hurts everyone, including myself, when we have to waste time refuting the bullshit and trying to find meaningful answers or solutions. I would say the answer is education, but short attention spans and the disparaging of concepts requiring more than 140 words to express suggest the problem is going to get worse, rather than better. And the very people who need to learn not to be fooled by these things are the ones who want to use this technology themselves to avoid having to do actual work. Welcome to the future.

Problem with LLMs (i.e. also ChatGPT) is that they lie in about 20% of the cases. They do it with a great confidence, good grammar and in an authoritative style. The result is that an LLM can fool you more easily than an average human. One must be aware of this when asking questions outside his/hers sphere of expertise. The lies are politely called "hallucinations" and they very likely cannot be fixed without extending the LLM model. These extensions are being done so there is chance of an improvement.

The problem with pure LLMs is that their reasoning is limited to a simple deductive specialization only.That is the only form of reasoning somewhat supported by the architecture directly. Everything else you get from an LLM is only a statistical mashup of the data it was trained with. It is essentially a statistical database and as such will sometimes generate wrong results.

LLMs are good source of proposals which are later verified and extended by humans. That is if you care about correctness. They are great even now to generate internet spam (mostly to support marketing and search engine optimization) where high error rate does not matter.

[webwit]

Still waiting for my flying car and voluptuous AI robot butler.

[lhutton]

Still waiting for my flying car and voluptuous AI robot butler.

Too bad, enjoy some weird six double jointed finger having person art and auto generated Amazon product descriptions instead!