padolsey

10 ms·

121.
by padolsey 1y ago
I wonder if it's just our egos talking, telling us that for something to have value within this technical sphere it has to be complex and 'hard won'? I see this agent stuff as a pretty basic agreement wrapped up in the vernac
122.
by padolsey 1y ago
HTML was designed prior to XML fwiw, spinning off from SGML.
123.
by padolsey 1y ago
I wish they had some example completions in the paper and not just eval results. It would be really useful to see if there are any emergent linguistic tilts to the newly diverse responses...
124.
Ask HN: Engineers and Academics: are you worried about entering the US?
4 points by padolsey 1y ago|3 comments
125.
by padolsey 1y ago
It always felt like a fun challenge. For me, it was a crucial part of my love of the web (and eventually, programming): the hackability and the way that there was no canonical correct way to solve a given problem. I miss those days.
126.
by padolsey 1y ago
Semantically speaking, tables are supposed to contain tabular data, not to be used solely for their layout capabilities. That's why it's called an abuse.
127.
by padolsey 2y ago
Well, the same frailties and socio-biological forces exist on all sides of the ocean. We are all lulled into entrenchments in defensive comfort, as are Americans. Right now, America is projecting an erratic personality and a chaotic power h
128.
by padolsey 2y ago
When recently visiting, what shocked me most about the CBP–which I presume is of a softer ilk than ICE–was the sense of entitlement and strength they projected, a kind of "prove to me your worth at the gates of our great nation" v
129.
by padolsey 2y ago
Agree! I wish Americans would travel more and escape their media bubble. China is incredible. I was in Chonqing and then Shenzhen recently, and live in Beijing. Truly it's very divorced from what western people have grown to expect.
130.
by padolsey 2y ago
This reminds me of how it's possible to decode prairie dog communication into literal 'words' that specify location of possible threats (lookup 'Slobodchikoff prarie dogs'). It's remarkable. Only made possible
131.
by padolsey 2y ago
I remember working for a small agency in London with a bit of a top-heavy management structure. Many times would my shoulder be subtly gazed over to "check on progress". I ended up using IRSSI to chat with friends, wgets to read m
132.
by padolsey 2y ago
What precisely is your point? There are various benefits not contingent on age, so while it is concerning, it by no means would by itself invalidate someone's entitlement in receiving the checks. There is probably some interface or was
133.
by padolsey 2y ago
Saddening. The gall to assume they can walk into a large institution with computing infrastructure older than them and start poking and pulling cables out without thinking for a moment. The arrogance it takes to not stop, to not doubt, to n
134.
by padolsey 2y ago
Is there a way to download and read this as a document instead of web pages? They're hard to navigate.
135.
by padolsey 2y ago
Perhaps the meaningful axis is not around seniority, per se, but instead around experimentation as OP alludes to. By virtue of being a junior, as you're just picking things up, you have to be flexible, malleable and open to new thing
136.
by padolsey 2y ago
> news exists for the sole purpose of creating outrage in order to generate ad impressions I like the idea of distinguishing news from journalism. If we say they're distinct, then yeh I think I can agree that news is–via weird unint
137.
by padolsey 2y ago
I'm personally ok with such rhetoric when the stakes are high (I am losing hope that "talking nicely/sweetly" to people is _actually_ beneficial, on net), though I think yeh the parent comment could do with more material
138.
by padolsey 2y ago
> Edit: decided I don't want to engage on this issue. I think I want that on my headstone :p
139.
by padolsey 2y ago
OOC What are examples of things the US is doing right now that you'd like Norway/EU to pursue? Like, specifically?
140.
by padolsey 2y ago
It's rare to find a subscription service doing something kind-hearted, sensible and good-faith towards their customers. There are so many dark-pattern subscription practices out there. Thank you - you've got a new signup :)
141.
by padolsey 2y ago
Many of these evals are quite easy to game. Often the actual evaluation part of benchmarking is left up to a good-faith actor, which was usually reasonable in academic settings less polluted by capital. AI labs, however, have disincentives
142.
by padolsey 2y ago
This seems to be a general pattern emerging. Cynical opportunists are wrapping hf endpoints/embeds in dodgy SaaS offerings. A similar one is BetterDictation, which tbf I do use. But I still hate that people are profiting off open-spiri
143.
by padolsey 2y ago
Gosh! Such inspired venom and "old man yells at cloud" vibes in this comment thread. I think I should send this to myself from 5 years ago: It's gonna be OKAY. It's OKAY to subtly adjust your language towards better ends
144.
by padolsey 2y ago
Tbf have you read most of the academic papers on LLMs lately? It's all a lot of boilerplate academic language packaging around "we tried prompting it this way and it did good things". Tho yes I do appreciate some scientific p
145.
by padolsey 2y ago
Cool! I love seeing the very subtle emergent behaviours of different CoT approaches. I reckon people still don't fully appreciate the brittle artistic subtlety of trying to make something akin to logic emerge from these weird transform
146.
by padolsey 2y ago
Now having worked in a few smaller teams since my time at FB, I realize how marvellous it is to not have to worry about OKRs, performance reviews and all that malarkey. It was so draining to run yourself as a mini PR/brand within one&#
147.
by padolsey 2y ago
Running code would be a downstream (client) concern. There's the ability to get structured data from LLMs (usually called 'tool use' or 'function calling') which is the first port of call. Then running it is usually
148.
by padolsey 2y ago
I think you mean this: https://arxiv.org/abs/2412.06769 From what I can see, presuming o3 is a progression of o1 and has good level of accountabiltiy bubbling up during 'inference' (i.e. "Thinking about
149.
by padolsey 2y ago
I agree; it has its part to play; I guess I just see it as such a miniscule one. True bad actors are going to use abliterated models. The main value I see in alignment of frontier LLMs is less bias and prejudice in their outputs. That'
150.
by padolsey 2y ago
> Even if it’s not a perfect solution, it help steer the problem in the right direction Yeh tbf I was a bit strong worded when I said "pointless". I agree that perfect is the enemy of good etc. And I'm very glad that they&

More ›