potatolicious

9 ms·

61.
by potatolicious 1y ago
Security is the big one. C++ DLLs have relatively free reign and are difficult to sandbox. With WASM you have a much stricter security model where the host program has full control over what APIs it has access to. The addons are developed b
62.
by potatolicious 1y ago
I think a few points worth making here: - From what I can see many products are rapidly getting past "just prompt engineering the base API". So even though a lot of these things were/are primitive, I don't think it'
63.
by potatolicious 1y ago
> "they can just come eat your lunch once it looks yummy enough. Tread carefully" True, and worse, they're hungry because it's increasingly seeming like "hosting LLMs and charging by the token" is not te
64.
by potatolicious 1y ago
Worth adding that sometimes I will spend an ~equivalent amount of time doing something in Claude Code, but the result is better. Case in point recently I was working on a mobile app where I had to check for a whole litany of user permission
65.
by potatolicious 1y ago
> "producing solutions for actual problems that real people have, is certainly intellectually demanding and also something that allows for several standard deviations in skill level" Absolutely agreed, but I think the idea is
66.
by potatolicious 1y ago
+1, and also to add: this isn't even a new phenomenon . My hot take is that AI coding is only the latest in a trend that has been running in the industry for literally decades. The technology itself might be novel but what it's d
67.
by potatolicious 1y ago
Yep, at that point we've circled back to the original years-long conversation about home servers, except now instead of a cheap mini-PC it's a phone. The distinction isn't meaningful. And I'll remind folks that we'v
68.
by potatolicious 1y ago
A whole host of reasons: - Battery life. One of the main reasons your phone lasts as long as it does is because it severely restricts the ability to run always-on things. A phone of course can run an email server, but the battery life wil
69.
by potatolicious 1y ago
> ” A very small model could run on device to automatically switch and choose the right model based on the request.” This is harder than it looks. A “router” model often has to be quite large to maintain routing accuracy, especially if
70.
by potatolicious 1y ago
AFAIK not for the program itself. Plenty of ability to page for other runtime memory consumption. I can only assume this is done with a presumption that responsiveness is a greater priority than the ability to accommodate extremely large bi
71.
by potatolicious 1y ago
A key tip that wasn't in the blog post: be extremely judicious about taking on new code dependencies. For many apps the main source of memory usage isn't stack or heap memory consumed during runtime, it's the loading of the b
72.
by potatolicious 1y ago
Ditto, it took me years to figure out "dynamic programming". It sounded hard and complicated! When I finally figured out wtf it meant I felt anger instead of relief - what an absolutely asinine name for what is otherwise a simple
73.
by potatolicious 1y ago
It is birthplace, so if you were a Bengladeshi national but were born in CH, you would fall into the Swiss queue (which is the "everyone else" queue). If this is you, congratulations, this is the shortest possible line. Though, if
74.
by potatolicious 1y ago
This feels like a philosophical fault line in the industry. For people whose purpose is to produce reliably working systems yeah, training a model that calls out to deterministic logic to do things like math makes total sense. It will prett
75.
by potatolicious 1y ago
> "I'm particularly annoyed by using LLMs to evaluate the output of LLMs." +1, and IMO part of a general trend where we're just not serious about making sure this shit works . Higher scores make stonks go up, who c
76.
by potatolicious 1y ago
Something like it definitely, though not exactly. We also know that recall improves with proximate position of bits within the context. Adherence to context is lossy in a way reminiscent of human behavior but also different in crucial way
77.
by potatolicious 1y ago
Yeah, I think it honestly lets founders off the hook too much. In 95% of cases, the founder isn't smashing moral barriers because the VCs and shareholders are making them, or lording the threat of legal action or any some such. In 95
78.
by potatolicious 1y ago
I take their point to be that the underlying incentives haven't changed. The same forces and incentives that scuttled those things are likely to scuttle this as well. I actually disagree with the OP in this sub-thread: > "No,
79.
by potatolicious 1y ago
Sure, and MCP is just a standardized way of exposing tools. This is where I feel MCP is both overhyped (waaaaaaay too much LinkedIn influencer hot air) but also genuinely quite useful. I've done stuff very much like the above with just
80.
by potatolicious 1y ago
> "on what timescale though?" Depends on how extensive the growth is and how structurally stable the thing is in the first place. But expect major problems on the decades-timescale. > "and in an invisible way?"
81.
by potatolicious 1y ago
> "You're claiming that your work is complex and unknowable to outsiders" I didn't get that message at all. If anything they're saying that the complexity of PM work is entirely knowable, but the many engineers
82.
by potatolicious 1y ago
I think this is at least somewhat true anecdotally. We do know that as context length increases, adherence to the system prompt decreases. Whether that de-adherence is reversion to the base model or not I'm not really qualified to say,
83.
by potatolicious 1y ago
A good point. The author does briefly address the point of mobile internet but I think it deserves a lot more real estate in any analysis like this. A few more points worth adding: - Depending on your product or use case, somewhere between
84.
by potatolicious 1y ago
That's kinda my main beef with Ed's writing - it's pretty unfocused. Both pieces you linked demonstrate this - he careens from point to point. The lack of focus I feel like comes from the fact that his audience is an amalgama
85.
by potatolicious 1y ago
I think there's some overlap in the thrust of both pieces but they're pretty distinct topic-wise? The first piece strikes me as a paean against "enshittification" - the idea that the industry has done good things, but
86.
by potatolicious 1y ago
Not to mention even just this article exposed a just-for-fun API that ended up having a negative effect and had to be removed: `isUserAGoat` ended up allowing any caller to determine if a specific app is installed on the system, which is a
87.
by potatolicious 1y ago
+1. I sometimes hear nostalgia for the N900 but personally I don't get it. Anybody could run a full multi-tasking OS on a mobile device trivially. The performance sucked and you killed your battery super quickly. The innovation was i
88.
by potatolicious 1y ago
So classic they still use it! iOS now offers a lot more multitasking options, but for the most part when you swipe away from an app it's still good ol' Mr. Screenshot. And if you'll excuse more nerding out - a lot of work is
89.
by potatolicious 1y ago
Agree. I'm sympathetic to the CTO here, but I remember the disaster of the HP TouchPad launch very well - there were multiple fatal errors here that don't seem possible to commit in an 8-week window. The hardware had basically ~no
90.
by potatolicious 1y ago
This is true - and WebOS was legitimately innovative here. At the time neither iOS nor Android could run more than one app at the same time. This was both an architectural matter and a UX matter. On iOS and Android at the time, all apps wer

More ›