girvo

13 ms·

211.
by girvo 5mo ago
The big question for me having used a lot of these SOTA chinese models is: what is its token efficiency like? Running Step 3.5 Flash locally for example, it's an amazingly capable model all things considered, but it's token effi
212.
by girvo 5mo ago
Their ability is best described as "spiky". To steal from aphyr: think kiki, more than bouba. Whats interesting is that a lot of the models seem to have similar spikes and "troughs", though there are differences.
213.
by girvo 5mo ago
As a dispassionate third party: your assertion is literally just as baseless unless you provide said base. It’s wild to shout down someone else when you yourself are doing the same thing.
214.
by girvo 5mo ago
I’m sort of surprised that they’re not banned already; dating myself but when I was at Uni in the late 2000s they were banned then. Despite probably not being very useful for cheating on nascent 3G!
215.
by girvo 5mo ago
It’s more commonly posted here these days, but this has always been on this site. Just usually couched “better”
216.
by girvo 5mo ago
Corporations don’t follow or care about morals, and so their customers and workers begin to follow suit.
217.
by girvo 5mo ago
Literal wage theft is rampant, so yeah in some ways I understand why people would feel that way. I’m still mad about a company I worked at over 12 years ago who stole from me and never paid my Super.
218.
by girvo 5mo ago
Which is amusing because “proving grounds” is still using the old definition of “proves” :)
219.
by girvo 5mo ago
Yeah that’s basically what I mean! I have no issues wrangling it myself, but now I’m curious how those who are managing “fleets” of agents while shipping four features a day are doing it. They’re not, I’d assume?
220.
by girvo 5mo ago
But… it doesn’t matter? Even if it was some very illegal drug, that doesn’t change the fact that this detention system (and Japans justice system in general) is quite inhumane.
221.
by girvo 5mo ago
I’ve struggled to get Opus to not write the weirdest possible Rust, ignoring all idioms and so on. Any tips?
222.
by girvo 6mo ago
Oh that's fascinating. 3.6 27B is pretty damned good, but slow in wall-clock times on my DGX Spark-alike. It generates huge reams of thinking before it gets the (usually correct!) answer, so wall-clock time is rough for tasks even at
223.
by girvo 6mo ago
I'm faster, sure, but more thorough, no. The same, because I was already very careful. But it's not a massive win either; 4.7 misses too much still because it would need to read too much of the context each time to understand the
224.
by girvo 6mo ago
You can, though. We used pthreads (well, pthresd compatible API) in production at massive scale on the ESP32-S3.
225.
by girvo 6mo ago
It’s been quite a few years since I last did PHP, but I wrote my own wedding invite management tool using Laravel and PHP 8.5 Herd is super neat, reminds me of XAMPP back in the day. FrankenPHP is everything I wanted out of a modern PHP web
226.
by girvo 6mo ago
The dirty secret is all the people talking about shipping 4 features a day etc are just lying about reviewing anything. They don’t review it at all.
227.
by girvo 6mo ago
Frenchisms are great! “No A, No B, No C. Just D” in every comment “assisted” by AI isn’t that great :(
228.
by girvo 6mo ago
It’s flagged because it’s obvious AI slop
229.
by girvo 6mo ago
I’m flagging them, but it’s not enough. Front page is full of it these days. Sad times.
230.
by girvo 6mo ago
Already sort of exists in the high end contracting/consulting software dev business in Australia at least
231.
by girvo 6mo ago
Through Pi, mostly! Also my own for-fun agent I wrote Yeah so would I, I do miss having vision tools sadly.
232.
by girvo 6mo ago
Of course, they’re shockingly powerful, just in an incredibly “spiky” way
233.
by girvo 6mo ago
The literal best public models still fail to count characters consistently in practice so I’m not sure what you mean. It’s literally a problem we’re still trying to solve at work
234.
by girvo 6mo ago
Yeah exactly. Shader model 121 is different to SM 120 (consumer Blackwell) and is different again to data centre Blackwell SM100. The promise of this chip was “write your code locally, then deploy to the same architecture in the data centre
235.
by girvo 6mo ago
> Does anyone really do this Yes, lots of people. It’s a whole issue.
236.
by girvo 6mo ago
> The claim you should know everything about everything you work on is an intensely naive one Nothing in the article made that claim.
237.
by girvo 6mo ago
True, though the $20 equivalent used for planning only I don’t hit those limits often, vs Claude where the Pro can literally hit limits with a single prompt haha
238.
by girvo 6mo ago
I heavily considered one of the AMD Strix Halo boxes, but part of the reason I wanted this was to learn CUDA :)
239.
by girvo 6mo ago
It's surprisingly good. Beats MiniMax 2.7 and Qwen 3.5 Plus in my testing (I haven't tested 3.6 plus though), quite handily. It's far better than Sonnet, and often equivalent to Opus for the web development and OCaml tasks
240.
by girvo 6mo ago
> DGX Spark-alike Probably wasn't clear enough if you don't know what that is already, apologies It's an Asus Ascent GX10, which is a little mini PC with 128GB of LPDDR5X as shared memory for an Nvidia GB10 "Blackwell

More ›