jeremyjh

16 ms·

361.
by jeremyjh 5mo ago
Most people only use a subset of language, and a lot of language is used only by a small subset of people. "The Ask" is in mainstream usage - or at least known to - the target audience of TFA, which may not include you.
362.
by jeremyjh 5mo ago
Are you educated? Descriptivism is overwhelmingly the mainstream view or practice of language scholars. https://en.wikipedia.org/wiki/Linguistic_description
363.
by jeremyjh 5mo ago
That is absolutely the only requirement for it to be correct. Words are defined by the way people use them, and that can change over time. A dictionary is a record, not a rulebook.
364.
by jeremyjh 5mo ago
"The Ask" is in widespread usage as a noun.
365.
by jeremyjh 5mo ago
Exactly this. GLM 5.1 is the first open model that I thought "actually worked" for agentic coding, which puts it in the same tier as Opus 4.5 - which was where I flipped.
366.
by jeremyjh 5mo ago
Yes people are speaking past each other to some extent here; but if we're going to talk about "Feature Flags" we aren't talking about using a flag service for "normal" configuration - whatever that means. But w
367.
by jeremyjh 5mo ago
This is profoundly misinformed. I use all three of those models regularly and the difference is just not that big anymore. GLM 5.1 is at least as good as Opus 4.5 - when it’s my dime it’s the primary model I use and switch to GPT 5.5 for pl
368.
by jeremyjh 5mo ago
One well known issue is that when you have a lot of separate feature flags that can interact, you explode the number of test cases you have to cover. For example if you have three feature flags that can interact in a module that has 100 tes
369.
by jeremyjh 5mo ago
Load-bearing slop.
370.
by jeremyjh 5mo ago
Its more about how costly it is to verify work in reinforcement learning. It is cheap in Mathematics and coding because it can be automated. It is expensive in other domains because while you can capture certain datasets to do pre-training
371.
by jeremyjh 5mo ago
Right, so you have to use RLHF. That is the economics problem I was referring to.
372.
by jeremyjh 5mo ago
The funders are among the principal builders in this context. This is addressing the people who have a say in what is built, and how it is built. Much of that belongs to the executive and ownership class, but not all of it.
373.
by jeremyjh 5mo ago
Its not surprising, that doc is full of AI slop.
374.
by jeremyjh 5mo ago
Yeah Linux will be shitty and break all the time if it doesn’t come preinstalled with Brave and a shortcut to open Grok.
375.
by jeremyjh 5mo ago
Even the strongest frontier model they used - GPT 5.2 - I would consider barely usable for agentic programming. I’m not really interested in analysis of the weaknesses of such models because in my experience many weaknesses disappear entire
376.
by jeremyjh 5mo ago
I think that’s mostly because they get so much more of that reinforcement learning - since it is so economical. I dont know if there is any evidence of a fundamental reason they can’t be just as good at other tasks, but it might be economic
377.
by jeremyjh 5mo ago
I had many similar experiences, but almost a decade earlier. At grade school we had Apple 2s with Logo, Oregon Trail and other education classics. My junior high was a small parochial school that still had TRS-80s in 1988, along with some a
378.
by jeremyjh 5mo ago
Maybe there should have been a plan. A period of notice and a transition plan. It couldn't happen, since the Trump administration does not believe in competence, only in spectacle.
379.
by jeremyjh 5mo ago
The Haskell community wrestled with this for many years - may still I haven't really kept up. Last I was involved the most successful thing going on was Stackage where they encouraged no defensive upper bounds and ran continous builds
380.
by jeremyjh 5mo ago
What kind of rules have been successful? Is it something that is constantly shifting and you have to react to, or WAF handles it based on usage patterns?
381.
by jeremyjh 5mo ago
I didn’t know his name but certainly knew about MATLAB. He sounds worthy of a black bar to me.
382.
by jeremyjh 5mo ago
I've been using Kagi for a couple of years now. For awhile, my work Chrome profile was still set to use Google for search by default because I don't login to personal accounts on it. I didn't use search that often for program
383.
by jeremyjh 5mo ago
> Someone's got to explain to me how this was even remotely plausible. You need to understand what a nuclear chain reaction is. https://en.wikipedia.org/wiki/Nuclear_chain_reaction > We've had orders of
384.
by jeremyjh 5mo ago
Stockfish did not teach itself to play chess. You are probably thinking of Leela Chess Zero - an open re-implementation of AlphaZero - both were given nothing but the rules of chess and a board and played millions of games against themselve
385.
by jeremyjh 5mo ago
> LLMs don't really create novel mathematics, they mostly "connect the dots". That is not what the mathematicians are saying. I don't have the knowledge to evaluate this myself, but a number of mathematicians - for e
386.
by jeremyjh 5mo ago
You just turn it all off. So yes, the disruption is the same but restoral is much smoother. Much easier said than done - that has be baked into every service and there would certainly be a cost from it that would have to be passed along to
387.
by jeremyjh 5mo ago
In many places in the world, the only interests of government are those of the rich. I do not think the rich will be petitioned so much as they will be … persuaded.
388.
by jeremyjh 5mo ago
They even used the distinction “native-like” in the block editor article - which is really good, by the way and explains this distinction in more depth - but edited their comment now and that article is the third link and its anchored to th
389.
by jeremyjh 5mo ago
> I know nothing about any of these APIs Agreed. In Chromium all the content from HTML is rendered inside a single object from the point of view of the host UI; much like a game engine’s UI rendering. Chromium draws everything itself. Ho
390.
by jeremyjh 5mo ago
You can imagine anything you want, but it’s not an argument - you could apply this to anything. “Python was successful after a dubious beginning so NFTs will be successful” Also, Python does not build or run large language models. It orches

More ›