Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
scubbo
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
13 ms
·
181.
▲
by
scubbo
1y ago
> I'm going through a bit of a phase at the moment, so I'm biased. It's "show me the incentive and I will show you the outcome". I've been in that "phase" for several years now. It's rarely st
182.
▲
by
scubbo
1y ago
They reference Veblen in the very first footnote.
183.
▲
by
scubbo
1y ago
I've had two people tell me in the last week that SQS doesn't support FIFO queues.
184.
▲
by
scubbo
1y ago
A barrier-to-entry isn't necessary if no non-geeks want to join in the first place (which is, IME, true of Mastodon). Mastodon has successfully made itself a place that no non-geek would ever _want_ to spend time on anyway, so no barri
185.
▲
by
scubbo
1y ago
a) GNU Terry Pratchett b) In case you are one of today's Lucky Ten Thousand, this is a reference to the real-life Ceremony Of The Keys[1] [0] https://xkcd.com/1053/ [1] https://www.hrp.org.uk/tower
186.
▲
by
scubbo
1y ago
Thank you!
187.
▲
by
scubbo
1y ago
What about Mastodon? (I'm, like, 80% joking)
188.
▲
by
scubbo
1y ago
Would RAG also be an approach here? My intuition from some small investigation is that RAG is more formal and structured to set up, but more efficient, whereas MCP you can just point an LLM at an MCP server and tell it to figure shit out (a
189.
▲
by
scubbo
1y ago
I mean this earnestly, not snidely - I wish I still had the faith that you do in not being treated abominably by any and every company, or to believe that they wouldn't default to behaving improprerly at any opportunity and for the bar
190.
▲
by
scubbo
1y ago
I think you're making good points, that aren't exactly counter-examples to the concerns being raised. You are making the - correct! - point that _other companies_ who have paid contracts with an AI provider would impose significan
191.
▲
by
scubbo
1y ago
I bounced away from Gitea because they don't (last time I checked) have OIDC. I started[0] trying to revive-and-drive a previous PR[1] to add it, but the test failures are beyond my motivation to investigate and resolve. [0] https:&#x
192.
▲
by
scubbo
1y ago
Increasingly realizing that I'm not cynical enough :'(
193.
▲
by
scubbo
1y ago
> I try to be mindful of what I share with ChatGPT, but even then, asking it to describe my family produced a response that was unsettling in its accuracy and depth. > Worse, after attempting to delete all chats and disable memory, I
194.
▲
by
scubbo
1y ago
> You really think everyone is actually sick on every single sick day, or that every single doctor's appointment on their calendar is a real doctor's appointment? ...I.....used to. Huh.
195.
▲
by
scubbo
1y ago
Nope, it's not! It's more than the bare minimum I need, and also larger than an appropriate moderate size (let's not shift the goalposts here) because - unlike a vehicle - operating a larger-than-appropriate home is _barely_
196.
▲
by
scubbo
1y ago
Because accomplishing the same task with a more powerful (i.e. larger) vehicle is a) more polluting, and b) more dangerous for other road users; two things that are not true for a surplus of computing power.
197.
▲
by
scubbo
1y ago
> [it's fine to expose your services to the Internet, there will be] no consequences. > and then Okta, which should be pinnacle of security gets hacked and has issues with disclosure. Pick one. If even the "big boys" ca
198.
▲
by
scubbo
1y ago
Read on in my comment for the more important point about fragmentation. Average non-technical users won't - and shouldn't have to - accept having to switch between different Tailnets to access different instances of similar apps h
199.
▲
by
scubbo
1y ago
> It comes off to me as the author not wanting to do the hard stuff of working towards their values Unfair IMO. The author _did_ the hard work. And recognized that most other people, not similarly motivated, would not.
200.
▲
by
scubbo
1y ago
> exposing our services to the public internet You yourself have hand-waved away an important part - security. It's not (just) about the friction of signup (though, I'll get to that later) - it's the fact that you'd b
201.
▲
by
scubbo
1y ago
Hurrah for do-nothing scripting[0]! [0] https://blog.danslimmon.com/2019/07/15/do-nothing-scripting-...
202.
▲
by
scubbo
1y ago
That would be true if I was making a argument of criticism of a certain person or class-of-people, but I'm not. I'm describing my observations about the frustrations that AI-skeptics feel when they are bombarded with contradictory
203.
▲
by
scubbo
1y ago
I have seen people suggesting that it's OK that our codebase doesn't support deterministically auto-adding the import statement of a newly-referenced class "because AI can predict it". I mean, sure, yes, it can. But dras
204.
▲
by
scubbo
1y ago
Yep, precisely my point. Reviewing a one-shot in 15 minutes is irresponsible.
205.
▲
by
scubbo
1y ago
With the exception of 1) being "No True Scotsman"-ish, this is all very fair - and if the technology was presented with this kind of grounded and realistic evaluation, there'd be a lot less hostility (IMO)!
206.
▲
by
scubbo
1y ago
> I was being sold a "self driving car" equivalent where you didn't even need a steering wheel for this thing, but I've slowly learned that I need to treat it like automatic cruise control with a little bit of lane sw
207.
▲
by
scubbo
1y ago
Eh, phrasing ambiguity from the run-on sentence (analogous to how "I fought with" could mean "on the same side as" or "against"). Yes, I meant that LLMs will do harm to - will exacerbate - _the problem of"
208.
▲
by
scubbo
1y ago
> Use the tools properly > (from upthread) I was being sold a "self driving car" equivalent where you didn't even need a steering wheel for this thing, but I've slowly learned that I need to treat it like automatic
209.
▲
by
scubbo
1y ago
> Hours of time saved Come back in a week and update us on how long you've spent debugging all the ways that the code was broken that you didn't notice in those 15 minutes. Usually I don't nitpick spelling, but "mimnu
210.
▲
by
scubbo
1y ago
> LLMs are cheap enough to run profitably on ads alone > It is even cheaper to serve an LLM answer than call a web search API These, uhhhh, these are some rather extraordinary claims. Got some extraordinary evidence to go along with t
More ›