jmward01

18 ms·

331.
Minnesota Bathrobe Lady Sam Stroozas of MPR News(mpr.org)
4 points by jmward01 8mo ago|1 comments
332.
by jmward01 8mo ago
Odd.. Not sure that cut-and-paste happened. For those interested the like I was going for was: https://mashable.com/article/homeland-security-issues-subpeo... The title changes Was Homeland security to DHS since the or
333.
by jmward01 8mo ago
Ugh! I hate cut and paste. I'll delete this since it has been done. Thanks for checking my work. [edit] The 'delete' option is gone. I guess there is a time limit on it? Sorry for the bad submission. I'll take my lumps a
334.
by jmward01 8mo ago
We have a problem right now there is a lot of a bad 'news' sites and the few that do any good get slammed because they listen. Go ahead, slam Fox news and see how far that goes. I think this creates a very negative incentive to be
335.
by jmward01 8mo ago
I see a lot of negative comments on this retraction about how they could have done it better. Things can always be done better but I think the important thing is that they did it at all. Too many 'news' outlets today just ignore t
336.
by jmward01 8mo ago
I totally agree. US healthcare is broken and costs aren't tied to the reality of how expensive something actually is. I have very high hopes that modern medicine is in for a massive disruptive change where things like full body MRI, al
337.
by jmward01 8mo ago
Maybe the right answer isn't to do a biopsy, but to monitor the area with follow-up scans? It seems like that addresses much of the harm that a false positive can cause (invasive biopsy leading to complications) while maintaining most
338.
by jmward01 8mo ago
Directly quoting Scott Shabaugh here: “Well if the code was good, then why didn’t you just merge it?” This is explained in the linked github well, but I’ll readdress it once here. Beyond matplotlib’s general policy to require a human in the
339.
by jmward01 8mo ago
What takes money out of their pockets is not paying a real wage for a real job. TIP destroys the value of a profession and when you don't think a profession is 'professional' you pay less for it. This is a terrible dark patte
340.
by jmward01 8mo ago
Having read the original post and the GH comments about why the PR was denied I was really impressed by their policy. It shows a real effort to develop their community.
341.
by jmward01 8mo ago
The core discussion wasn't about the PR it was about the hit piece that the bot created outside of the repo. The original post talked about bot submissions being a normal thing and how they have, I think, a very reasonable approach to
342.
by jmward01 8mo ago
Children's brains grow faster than their bodies, I think, because if it was the other way around silly kid games would be really dangerous. These tools, unfortunately, are getting outsized abilities before the intelligence behind them
343.
by jmward01 8mo ago
I did a lot of text cleaning a while ago and we tried to normalize curse word spelling as part of that. That was, by far, the most interesting text cleaning I have ever done. It is really clear how much innovation in the English language is
344.
by jmward01 9mo ago
Do they respect my data? Why do they get to track me across sites when I clearly don't want them to but someone can't scrape their data when they don't want them to. Why should big companies get the pass but individuals not?
345.
by jmward01 9mo ago
I hope you found a place that feels like home!
346.
by jmward01 9mo ago
I predict a specific location then put a window around that. Of course you can predict a different location per head or multiple window locations per head as well. The cost is negligible (single linear embx1 size) so attn becomes a fixed co
347.
by jmward01 9mo ago
I'm trying to figure the same thing out for my stuff. I figured out a simple way to train location prediction so I'm using it for guided window prediction which is great for attn (predict a distance in the past to look at) and for
348.
by jmward01 9mo ago
I'll give them all a look. Thanks!
349.
by jmward01 9mo ago
To get a similar token/sec in training though you would need to swap batch size and seq length so you could have the massive batch size but then won't you start hitting memory issues with any reasonable sequence length? You would
350.
by jmward01 9mo ago
RNNs have two huge issues: - long context. Recurrence degrades the signal for the same reason that 'deep' nn architectures don't go much past 3-4 layers before you need residual connections and the like - (this is the big one
351.
by jmward01 9mo ago
I agree the problem has more than one aspect, but we made real progress that we are now backtracking on nationally. Local problems will always be local. Go to a <insert industrial town here> town and they will likely be pro that indus
352.
by jmward01 9mo ago
I built guided window attn (literally predict the position of the window) a while ago and that works great. Why are we still stuck on any form of attn that looks at the entire context in any meaningful way? Do humans work this way? Do I nee
353.
by jmward01 9mo ago
Points taken. After seeing this and other similar posts I suspect my thought experiment would fail and the areas would be highly polluted. Maybe instead we need pollution parks? The big point is we need some sort of constant reminder of how
354.
by jmward01 9mo ago
But their CEOs don't live next to the factories. My, completely impossible, thought experiment would have all management living near the plant. I bet if the CEO lived with the water near the factory they would make sure it was clean.
355.
by jmward01 9mo ago
yep. If any city embodies the strode it is LA.
356.
by jmward01 9mo ago
I remember going to LA in the late 80's and my eyes watering (I also remember the pants-less man on the side of the strode but that is a different story). Environmental regulations are a win. Unfortunately there is a large segment of t
357.
by jmward01 9mo ago
From your description it sounds like you have the most important stuff: variety, a willingness to do things and a willingness to seek advice. I doubt anyone on HN really knows what it takes to learn to be a coder in the new vibe world. It i
358.
by jmward01 9mo ago
I did rally simple frequency analysis based on corpus source a while ago and the results were super clear, you can tell a corpus by its frequency fingerprint. I wonder if something similar to this could fingerprint bot accounts?
359.
by jmward01 9mo ago
I've been doing this the last year (ish) and my battery lasts so much longer. I suspect my data usage has been slashed as well. I recommend everyone starts using airplane mode. I just wish I could put individual apps in airplane mode s
360.
by jmward01 9mo ago
I'm not sure if you are intentionally missing the point or not. I'll try one last time. I have very clearly not said 'this shouldn't be done'. I haven't said 'we shouldn't pay for social services or r

More ›