tech_ken

13 ms·

121.
by tech_ken 1y ago
Wow that's super cool, thanks for explaining! I will definitely keep an eye out; having that level of certifiability and replayability at a pipeline level is something my team has been having too much angst about. It would be incredibl
122.
by tech_ken 1y ago
> For what it’s worth, could possibly have an intermediate Python model that...materializes the results to a table I think this is kind of the answer I was looking for, and in other systems I've actually manually implemented things
123.
by tech_ken 1y ago
The Git-like approach to data versioning seems really promising to me, but I'm wondering what those merge operations are expected to look like in practice. In a coding environment, I'd review the PR basically line-by-line to che
124.
by tech_ken 2y ago
Every paper I have ever seen on the subject has concluded that controlling for SES explains racial variation in crime rates. This paper touched on it in one location but otherwise didn’t bring it up, so that’s the part I mentioned here. Fee
125.
by tech_ken 2y ago
Are you familar with the concept of "omitted variable bias"? This paper only takes it up once (AFAICT), but there's a pretty critical variable that they otherwise generally ignore: > Once socioeconomic status was considere
126.
by tech_ken 2y ago
I think this is a pretty terrible litmus test for the "okayness" of a statement. There are plenty of reasons one word might make lots of sense in a statement, while another would absolutely not. For example: estimates of the preva
127.
by tech_ken 2y ago
The point of data presentation is to gist the most salient trends; an interactive chart where you can zoom in to the lowest granularity of the data basically defeats the purpose of the plot in the first place. Similarly most animation in ch
128.
by tech_ken 2y ago
Mathematics is, IMO, not the axioms, proofs, or theorems. It's the human process of organizing these things into conceptual taxonomies that appeal to what is ultimately an aesthetic sensibility (what "makes sense"), updating
129.
by tech_ken 2y ago
> Or more generally, how do you verify a proof that you don't understand? This is the big question! Computer-aided proof has been around forever. AI seems like just another tool from that box. Albeit one that has the potential to pr
130.
by tech_ken 2y ago
Why does other peoples' sincerity or lack thereof dictate what I am allowed to be outraged by? "Whatabout whatabout whatabout"; what about you worry less whether other people meet your standards for legitimate outrage and wor
131.
by tech_ken 2y ago
> What is the point of all of this? Just my opinion, but the most obvious motives seem to be: * Breaking the back of the institutional opposition Trump experienced in his previous term * Flexing strength and creating a narrative of unita
132.
by tech_ken 2y ago
> I.e, in 2017 Equifax leaked data on 150 million people and no one cared (you get a free 6month credit check) What are you even talking about? People (myself included) were fucking livid! The reason we got the 6mo credit check was becau
133.
by tech_ken 2y ago
One wrinkle here is the choice of signed error instead of eg. RMSE; absolute error metrics don't care about rounding vs. flooring. Since the conclusion of the analysis hinges on this choice of error metric, I think some justification o
134.
by tech_ken 2y ago
I think it's about evaluating the practical strengths and weaknesses of genAI for coding tasks, and trying to pair your skillset (or areas of potentially quick skill learning) with the weaknesses. Try using the tools and see what you l
135.
by tech_ken 2y ago
Why would you spoil match 14 for me like this T_T
136.
by tech_ken 2y ago
> It’s ridiculously hard to find the optimal combination of prompts and examples and fine-tuning, etc. It’s a very large space, there are no easy abstractions to allow you to search through the space, LLMs are unpredictable and fragile,
137.
by tech_ken 2y ago
> Big companies are like government bureaucracies now Large organizations are, by definition, bureaucracies regardless of being public or private. Big companies were bureaucracies 100 years, and they will be bureaucracies in 100 years. I
138.
by tech_ken 2y ago
> For example, customers arrive at the grocery store in clusters due to traffic lights, schools getting out, etc. You're kind of just describing seasonality components and exogenous regressors; RNNs do actually function quite well f
139.
by tech_ken 2y ago
Second time in the current millennium is probably the talking point you saw; Reagan was the the last Rep president to win two terms with a popular vote majority.
140.
by tech_ken 2y ago
Also federal government will continue to weaken under Trump. Its primary domestic power is by acting as a big hose of money, if that dries up then what does it matter? EPA, NLRB, all the executive offices are basically gone already anyways,
141.
by tech_ken 2y ago
That happened like 6 years ago, now it's just a question of high score
142.
by tech_ken 2y ago
"Thinking token" is an interesting concept, is there more literature on that?
143.
by tech_ken 2y ago
> sick to tech and information freedom Do you not understand that the tools the government uses to control non-citizens are turned inward all the time? Military tech is tech, surveillance tech is tech. They are not chiming in on 'im
144.
by tech_ken 2y ago
Seems more like a 'compression' of the large number of games, or even like an approximate 'index' of the database
145.
by tech_ken 2y ago
> On a different note, there's an argument dating back to Hobbes' Leviathan that the State and corporations are in fact a form of AI Woah neat I've been wondering about this myself, any reading you'd recommend?
146.
by tech_ken 2y ago
This is not an education issue, conspiracy theory prevalence is driven entirely by social dynamics and motivated reasoning. Watch that 2018 documentary about flat-earthers; one of the main 'characters' is like an aeronautics engin
147.
by tech_ken 2y ago
The points about the effects of noise are super interesting. Kind of mind blowing to think about the sensitivity of our perception being so different across visual channels (color, shape, movement, etc).
148.
by tech_ken 2y ago
> assume there was no p-hacking Agree that this is definitely an assumption one needs to make, could easily be that BDNF was one variable among many unreported ones, and this case would be consistent with the other outcome variables in t
149.
by tech_ken 2y ago
Local-first generation wants big interconnection transmission grids (for moving excess peak generation to off-peak timezones), and the federal government is way better capitalized to support these kinds of projects. State or even sub-State
150.
by tech_ken 2y ago
Definitely not the first, surprisingly hard to find clear data about the endpoints of the ties themselves but this random court document has a fairly clear diagram on page 3: https://www.nrg.com/assets/documents/en

More ›