rfgplk

7 ms·

91.
by rfgplk 2mo ago
If OpenAI were actually serious about capturing marketshare they'd release an unlimited subscription plan. Rate limit users if you need to, but let people use their agents 24/7. Even at $1000/month most would take it.
92.
by rfgplk 2mo ago
> which impose a relatively high bar on what content they accept. Nope. In fact they accept utter garbage most of the time. I can't count the numbers of times I've caught pure errors (literal logic errors, code errors, mathema
93.
by rfgplk 2mo ago
Which is ironic, because self-publishing then should automatically qualify you for a PhD. In fact, I believe more people should take this approach. As far as I'm aware, there's no law or regulation from declaring yourself a "
94.
by rfgplk 2mo ago
> I think the world would be better off if the universities actively discouraged the idea of starting companies. This is the case in Europe and the entire economy is a flaming pile of garbage.
95.
by rfgplk 2mo ago
Ridiculous frankly.
96.
by rfgplk 2mo ago
I found this. https://www.hesa.ac.uk/news/29-08-2024/higher-education-stud...
97.
by rfgplk 2mo ago
> The reality is that an incredibly small minority of companies in the world do any real training or optimisation. It's unnecessary and inefficient for most purposes unless you are fully dedicated to being an LLM company, and still
98.
by rfgplk 2mo ago
Do you hold the same opinion with regards to Jonathan Blows jai language?
99.
by rfgplk 2mo ago
> And here I sit with my own native cross-platform GUI library, made with my own Lisp-To-Rust programming language... Currently working on something like this. Do you have a writeup somewhere? Specifically how you solved cross platform r
100.
by rfgplk 2mo ago
Lots to criticize here. > This might sound obvious, but it is worth putting it down, Software Development going forward will be largely done by AIs, you might find the quality subpar, but in terms of cost ratio, it is commercially good e
101.
by rfgplk 2mo ago
> Are you a physician? This sentence spoken on an engineering forum is pure... idiocy. Physicians are nothing more than bio-engineers, and if you apply sufficient reasoning you'll see that there is very little functional difference
102.
by rfgplk 2mo ago
Too extreme of a take. They're ignorant at best. In fact after cholecalciferol/calcitriol and the whole vitamin D receptor family, ignorance of the importance of bacterial balance and mineral intake are the next two biggest catego
103.
by rfgplk 2mo ago
> 5000 IU/day Ironically these "IU/day" estimates are bogus. Direct ~noon sunlight at the equator for fair skinned individuals produces (full body exposure) 15-50k IU/hour of D3 equivalent, depending on the estim
104.
by rfgplk 2mo ago
> Doctors are in fact notoriously bad when it comes to vitamin D. Most doctors know nothing about the endocrine system, at all. I self taught myself through "Endocrinology: Adult and Pediatric" and I probably know more than 99%
105.
by rfgplk 2mo ago
In their current form open weight models are simply not worth running. Literally the amortized cost of hardware + electricity you need to operate them is >> than the cost of paying for subscriptions.
106.
by rfgplk 2mo ago
They loosened the refusals recently. They aren't anywhere close to the Sol refusals.
107.
by rfgplk 2mo ago
> I've been not just unimpressed by Fable, but actively find it to generate negative value. Fable is good in a few very specific domains (graphics programming) but otherwise it's an overhyped model. Far too expensive too. Opus
108.
by rfgplk 2mo ago
> Almost 10 models are passing the benchmark >95% - isn't that... substantially overly saturated? The only true benchmark for any of these models that I've discovered isn't if they can pass precanned SWE tests, but rath
109.
by rfgplk 2mo ago
> If I need to be a domain expert anyway, the value of the tool goes down by orders of magnitude. Same if I need to first break the task down into pieces and keep reviewing all the output. That sounds to me like >80% of the work I&#x2
110.
by rfgplk 2mo ago
> If I can make up a figure, I would guess 95% of software engineering quality rhetoric, and craftmanship advice, is just strongly held opinions. Absolutely, at the end of the day the only metrics that matter are performance and code val
111.
by rfgplk 2mo ago
> I believe this because every time I use AI for domains that I consider myself above competent, if it is anything beyond UI components or a simple CRUD endpoints, I cringe at the quality of what it generates. In a long run session Fable
112.
by rfgplk 2mo ago
> If you are playing with a gun, it goes off and hurts someone - you are responsible despite intent. No, because LLMs are autonomous. To make your analogy more accurate, it's as if you had a gun that itself was free to decide who it
113.
by rfgplk 2mo ago
> Who gets prosecuted? No one.
114.
by rfgplk 2mo ago
You can formally prove the correctness of even massive programs.
115.
by rfgplk 2mo ago
Funnily enough, since most JS frameworks are utter junk, developers are better off instructing LLMs to write pure JS instead. Fun story, I was writing some 3D graphics animations for a website, and the only dependency I allowed was three.js
116.
by rfgplk 2mo ago
> This idea is the reason why software will keep getting slower and less reliable. Because it's wrong and yet people believe it. Not even close to being true. Prior to the advent of LLMs writing good SIMD was hard. Writing good SIMD
117.
by rfgplk 2mo ago
LLMs have essentially made any language above the C++/Rust systems level tier obsolete. The only reason you picked a language like C# or Python is often out of convenience of having a broad ecosystem or due to the lower level languages
118.
by rfgplk 2mo ago
Wrong. Memorizing without understanding is pointless. And I'd argue that rote memorization does nothing to guide ones understanding. Because if you actually understand a subject deeply enough you can literally derive almost anything th
119.
by rfgplk 2mo ago
> What is a doctor? The person recommending next course of action based on best practices. What is the best practice? Is it a globally decided on thing, a locally tested thing that has proven to be better? Is the global thing just the mo
120.
by rfgplk 2mo ago
A good analogy to caching and memory lookups. You store in your head only the most valuable immediately necessary information, the rest can be outsourced.

More ›