122 karma · joined June 16, 2014
Why do you believe this?
Very odd that people who want thicker skin in our culture would find it so distasteful that I've simply shared some recent history and some first-hand personal experience without any particular name calling or inflammatory remarks. Can't a person share their thoughts anymore? Really didn't mean to offend anybody and somewhat surprised that I have seemed to.
The justice system is ideally intended to provide fair and impartial justice, but I think it's also fair and reasonable to point out when it falls short of that ideal. In fact, in a democratic society, I would argue that an informed and vocal public is a necessary component of keeping institutions accountable to the people.
How many firings for presumed inflammatory conduct have you witnessed or hear from first-hand witnesses? Personally, I've been fortunate not to have experienced any of these, but I do believe they exist and reading about them is useful for informing my worldview. It's helpful to understand how corporations and institutions are ingesting and digesting societal trends around evolving cultural awareness of injustice in order to protect their commercial interests.
Very serious injustices have occurred within the recent past. Take, for example the 1985 MOVE Bombing, in which Philadelphia police bombed a house in a predominantly black neighborhood during an armed standoff with the MOVE group. Whether or not you feel the bombing of an armed group was a justified use of force by police, the subsequent 61 homes that were allowed to burn down are harder to justify. This was followed by the ethically fraught decision by the Philadelphia Health Commissioner to cremate/dispose of any human remains without contacting family members, with the ultimate result that those remains were used in UPenn and Princeton "forensic anthropology" courses without any chance for their families to reclaim their remains.
More recently, I personally witnessed Philadelphia police corral Black Lives Matter protesters into an enclosed space on the side of the highway with no exits and fire tear gas into the crowd. On the news, I watched a number of similar confrontations take place in multiple cities in the nation.
When inequality continues, I have come to appreciate that the ability to feel that a grievance is historical is a privilege. A significant motivation for people to turn to history is to better understand the struggles through which they are currently living. For these people, understanding the historical context is not a way to bring up old grievances, it is a lens through which they can properly understand how systematic disparities in due process and access have produced today's injustices. It is an aid to better identify how currently extant systems create unfair conditions at a large scale.
HN often can be a great place to seek interesting and varied discussion on purely technical topics. When discussions veer anywhere close to topics that may touch on how people want to view themselves and their relationship to the world, PEOPLE can be somewhat lazy, tribal, and narrow-minded and PEOPLE are bad at respecting other people's value systems that differ from their own: because commenters on HN are people, this includes discussion on this site. Intersectionality is poorly understood, "hurt people hurt people", displacement[2] is rampant and that can lead to some pretty toxic interactions. It takes a very low proportion of badly-behaved people to produce an overwhelming number of negative, hateful comments. I overall like the community here, but that doesn't mean that it can't produce bad behavior that is beyond the scale of an individual to cope with reasonably, and I can sympathize with an individual seeking to cut off ties with a source of such conflict. As a society, we still haven't really adopted to the scale of contact that the internet creates.
The [insert social networking site here] effect is real and can cause enormous traffic spikes, which have real bandwidth costs and in this case potential emotional costs; removing a tool to deal with that spike isn't exactly a neighborly thing for HN to do, even if I can understand other arguments promoting less tracking, the realities of the internet etc. that people may want to make.
[1] Displacement: an immature defense mechanism in which a person acts out their frustration and negative feelings on a target that is not responsible for the situation evoking those negative feelings. Typically that target is less powerful and less able to retaliate within a given setting. Colloquially "punching down". The specific populations that are marginalized/less powerful are very context-dependent. Classic example is Person A getting a bad review from their boss and yelling at their domestic partner. Punching down at that individual/population puts Person A in a position of power, gives them a feeling of agency and control, and provides a feeling of moral/commercial/etc superiority that they feel is lacking in other aspects of their life, but it comes at the expense of others.
> obviously if vile comments frequently come with an HN referrer, a site admin who reads the vile comments will think poorly of HN
You can hate the site and the moderation team without believing that everyone on the site is bad. You can recognize when a referrer is generating a lot of stress and heartache and take steps to reduce that, and you can be upset when your attempts to reduce that are circumvented by site ownership, all without believing that the majority of the folks using the site are bad.
The top right of the PDF-alike has a link that says "what's this". and it sends me to https://www.springernature.com/gp/researchers/sharedit which seems like a service that allows sharing of links to fulltext. If I go to the url directly, e.g. by copy-pasting the URL or clicking on your link, then I do not get the same behavior. Hence, I deduce that it is probably some referrer magic.
When I click the same BBC link on chrome on android mobile, I don't get the redirect, so probably there is some User Agent stuff too.
> Artificial intelligence (AI) and machine learning are often used interchangeably, but machine learning is a subset of the broader category of AI.
and
> Machine learning is a pathway to artificial intelligence. This subcategory of AI uses algorithms to automatically learn insights and recognize patterns from data, applying that learning to make increasingly better decisions.
From the article:
> Curiously, this experimental antibiotic had no effect on other species of bacteria, and works only on A. baumannii.
This is an oversimplified claim made by the BBC article, but broadly aligns with the research paper claim that the identified drug is specifically not broad-spectrum and not active against e.g. Pseudomonal and Staphylococcal species. The identified drug candidate is unlikely to ever see broad use because of the low activity against many clinically relevant bacteria.
The most common use case for this candidate, should it see approval through the FDA process for this indication, will be for treatment of hospitalized patients who, while waiting for culture results, have failed broad-spectrum regimes and whose culture results demonstrate Acinetobacter infections.
> The AI was then unleashed on a list of 6,680 compounds whose effectiveness was unknown. The results - published in Nature Chemical Biology - showed it took the AI an hour and a half to produce a shortlist.
"published in Nature Chemical Biology" is a link you have to click to see the article in fulltext, which I would encourage you to read if you really want to understand the study. I would link it directly, but there is some site-referrer magic happening that allows the BBC article link to cause Nature publishing group to show the fulltext.
To better understand what was done:
Training set (manually tested): off-patent drugs (2,341 molecules) and synthetic chemicals (5,343 molecules). In particular the synthetic chemicals are likely to have unacceptable side effect profiles. Result of manual screen: 480 molecules capable of inhibiting Acinetobacter growth by 20%. 480/(2341+5343) = 6-7%
Result of AI processing and additional filtering criteria: model applied to "Drug Repurposing Hub" dataset consisting of 6,680 molecules which they claim have demonstrably favorable cytotoxicity profiles and drug-like properties and yielding 3 sets of 240 drugs each. The three sets:
1. 240 drugs identified by the model as having >20% probability of at least 20% growth inhibition of Acinetobacter and structurally _dissimilar_ to those with antibiotic activity in the training set. Manually testing for those capable of more stringent criteria (80% inhibition of growth) yielded 9 drugs.
2. 240 drugs with lowest prediction scores: Manual testing yields no active drugs, providing some basic validation that classifier is functional.
3. 240 drugs with highest prediction scores (without additional filtering criteria based on structural dissimilarity): Manual testing yields 40 drugs capable of 80% inhibition. 40/240 = 16-17%. Yield enrichment: 16%/6% = 260% or a 2.6x improvement compared to naive screen of training set. To be fair, this isn't a direct claim of the paper for good reason: the drugs in their training set and "validation set/drug repurposing hub" are fundamentally different and may have different baseline antibiotic activity across the set.
The process of narrowing down these datasets using the model could be accomplished in hours (their claim) instead of days (my claim). Days is optimistic prediction, requiring high-throughput systems and/or staffing in place to run these screening assays mostly in parallel instead of serially. Also prevents costs associated with further biochemical investigation (cultures, chemical synthesis and assays are not free).
Most direct value of this work is in accelerating drug screening process, reducing cost and developing AI-tractable representations of pharmaceutically-relevant chemical features. Additionally, proof of concept for identifying drugs with appropriate side effect profiles that happen to have antibiotic activity but would not have been identified with existing/common structural analysis approaches, since they are structurally dissimilar to the testing dataset screen. In this case they identified a "CCR2− selective chemokine receptor antagonist" that had antibiotic properties; some googling suggests that this drug class mostly has roles in fibrosis/inflammation regulation and may have roles in autoimmune disorders and those with significant fibrosis as part of the pathology (e.g. cardiovascular disease, liver disease, diabetes). You wouldn't expect most drugs in this class to have any antibiotic properties and many companies would not focus their first efforts on screening such drugs with biochemical assays.
> Tough to penalize us because Android does not expose RCS APIs
As someone speaking from a genuine place of ignorance, the Google page relating to RCS https://jibe.google.com/ seems to imply that RCS is just a universal specification and there are a number of documents that seem relevant after a search for "gsma rcs specification".
Is this an Android permissions thing where the only practical way to implement RCS support would be through a Google-supplied API?
> Hispanic men, 25-29 years old, 6'0"-6'1". Filtered out by 9% of possible matches
This was a misquote. Actually instead of saying "Filtered out by 9%", the article says "Filtered out 9%". So Hispanic men in this age and height bracket created height filters that reduced the size of the overall candidates available to them by 9% of the population. It is NOT that they are rejected by 9% of the population.
Correcting the above 2 posts based on this new understanding:
============================================================
"[Specific Filter]" Observations ============
1. First option Age: Among those aged 30 or older, age filters are more restrictive among women than men. Men aged 35-49 have the most permissive age filters, with my guess being because they are more willing to see matches from younger women. Among those aged 20-24, men set up age filters that reduce the candidate pool more than the age filters set up by women in the same age bracket.
2. First option Height: Among those 5'6" or higher, women set up much more restrictive height filters than do men. This is not surprising and corresponds to the observation that women often want a man who is not a lot shorter than they are. Among those 5'5" and shorter, men set up more restrictive height filters than do women. This makes sense under a general understanding that short men may feel emasculated by dating a much taller woman, or alternatively that they would be comfortable in such a relationship but are trying to improve the success rate of their matches since their experience is they will be rejected by much taller women.
3. First option Race: For every race but one, women set up more restrictive racial dating filters than do men. I was somewhat surprised about this, and that black women apparently set up comparatively restrictive racial filters when compared to any other racial/gender demographic. Note that this restrictiveness reduces when you look at "All filters", suggesting that black women are therefore comparatively less restrictive in their filters with respect to age/height. The one place where this flips is Asians, which has a completely different implication than what I had thought under the old misunderstanding of the data. This means that Asian men set up more restrictive race filters than do Asian women. Perhaps this is a strategy to attempt to increase their successes with people who do get shown their profile? I.e. a similar approach to that taken by men 5'5" and shorter?
============================================================
"All Filters" Observations ============
1. filter out rates rise significantly compared to just looking at a single Age Filter/Race Filter/Height filter. As before, this makes sense but the implications are different following my updated understanding of the graphs. If you add the effect of more filters, of course you reject more of the potential population. What changes is the implication of the observation that the change between "[specific] filter" and "all filters" is most apparent for height and race. This means that actually people do not set up very restrictive age filters, but do set up more restrictive height/race filters. The "restrictiveness" of the age filters may have more to do with the age demographics of the population using League.
2. Show by Race: Race does not have a very strong effect on the distribution of filter restrictiveness, but there are some exceptions. For example, among men, Indian men appear to set up the most restrictive filters and Hispanic men appear to set up the least restrictive filters.
3. Show by Height: This basically shows the same relationship as for the "Height Filter" graph. Comparing the "Height Filter" and the "All Filters" version, two interesting results stand out. Men from 5'2"-5'5" have overall filters set up that produce similar levels of restriction as taller men. This suggests that men in this height range are attempting to broaden their dating parameters in race/age to "make up" for their height disadvantage. Men 5'1" and under may be doing the same thing, but still cut out a much larger portion of the population than their taller peers. Women 6'2" move very little on this graph when you switch "Height Filter" and "All Filter". This means that most of their "filtering power" can be accounted for by their height filter and they may have relatively broader race/age filters.
4. Show by Age: Men from 35-49 have the least restrictive filters overall. This isn't that surprising. Among women, those aged from 25-34 have the least restrictive filters. Also not very surprising.
Having said that, a quick google for any comparable data set doesn't seem available. So while the limitations are important to keep in mind, for now, this is the only data set that I'm aware of. My approach is to take it with a grain of salt, but not to completely ignore it.
Make sure to apply "All Filters".
Observations ============
1. filter out rates rise significantly compared to just looking at a single Age Filter/Race Filter/Height filter. This makes sense and the change between "[specific] filter" and "all filters" is most apparent for height and race. My intuition is that people are more likely to set up age filters than other types of filters.
2. Show by Race: Men are filtered out less frequently than women for every race. There are way more white users of the app than any other ethnic group.
3. Show by Height: Men don't get filtered out more than women until they are 5'1" and under. Apparently, as a 5'3" man you are less likely to be filtered out than a 5'3" woman. This is very surprising to me.
4. Show by Age: Men from 35-49 are the least likely overall to be filtered out. For every age group, you are more likely to be filtered out as a woman than as a man. This is again surprising to me, especially in the younger brackets.
[edit:formatting/linebreaks]
Change the "Show by" filter and make sure the second option changes to the corresponding Age/Height/Race filter
Observations ============
1. First option Age: If you are 29 or younger, you are more likely to be filtered out if you are a man than a woman. If you are 30 or older, you are more likely to be filtered out if you are a woman than a man. This corresponds to my expectations.
2. First option Height: Similar, with the cutoff being 5'5" or 5'6". This roughly makes sense I guess, but I would have thought that women 5'8"-5'9" would be less prone to being filtered out on the first pass.
3. First option Race: For every race but one, you are more likely to be filtered out as a woman than as a man. This was surprising to me. Unsurprising is that the one category this flips is Asian - Asian women are less likely to be filtered out, Asian men are more likely to be filtered out. "Everyone" is open to dating White, Other, Hispanic, None Given, with relative penalties for being Black, Indian, Asian.
[edited for clarity and to better reflect the graphs I was looking at]
___
You have to look at it on the Economist and not the archive.vn mirror. Y axis is Height (or whatever you've set in the "Show by drop down" and X axis is the title of the graph, "Share of possible matches removed via filters, %". Hover over a point and the graph becomes much easier to understand.
E.g. one point shows
> Hispanic men, 25-29 years old, 6'0"-6'1". Filtered out by 9% of possible matches
Each of the circles represents a group that is defined by a permutation of Race/Gender/Age/Height. I think size of circle represents size of that population in the dataset. In the case of my example, the circle for "Hispanic men, 25-20 years old, 6ft-6ft1in" was small, so there probably aren't many men fitting that specific combination in the data set. You also see intuitively that the circles for men 5'5" or shorter are small and the circles for women 6'0" or taller are smaller, corresponding to our intuitions that men that short or women that tall are less common in the overall population.
Y axis is Height (or whatever you've set in the "Show by drop down" and X axis is the title of the graph, "Share of possible matches removed via filters, %". Hover over a point and the graph becomes much easier to understand.
E.g. one point shows
> Hispanic men, 25-29 years old, 6'0"-6'1". Filtered out by 9% of possible matches
Each of the circles represents a group that is defined by a permutation of Race/Gender/Age/Height. I think size of circle represents size of that population in the dataset. In the case of my example, the circle for "Hispanic men, 25-20 years old, 6ft-6ft1in" was small, so there probably aren't many men fitting that specific combination in the data set. You also see intuitively that the circles for men 5'5" or shorter are small and the circles for women 6'0" or taller are smaller, corresponding to our intuitions that men that short or women that tall are less common in the overall population.
I need to figure out if this dataset is available somewhere for analysis because an equivalence of 11 inches of height is somewhat shocking to me. It's a much bigger effect size than I expected, but maybe that just means I have to update my worldview. It may also have to do with the demographics of the daters interested in women using the League app.
Also, if I'm reading the first graph correctly, the penalty at the filtering stage for being a women 5'6"-5'7" or taller is worse than being a man between 5'2" - 5'3". Just not what I expected.
>sorry for the big words? maybe look them up?
It's not that I don't understand the motte and bailey fallacy, it's that you are misapplying when discussing my posts. In a motte and bailey, the person making an argument would like to advance controversial claim B, but when B is attacked, they will defend claim A while continuing to advance or believe claim B. A is generally a non-controversial and trivial statement. What do you believe is the motte, and what do you believe is the bailey in my argument?
I have chosen at times to focus on what I believe is the main issue, namely that there is an ethical question around inclusivity and the ways in which it shapes access to and development of future technology. This is the argument I am advancing, and one that is informed by a compassionate desire to believe people who say that something makes them feel excluded. If I choose not to be derailed by quibbles over technicalities of categorization that do not change the fundamental nature of my claim, that is not a motte and bailey.
A more correct use of motte and bailey could be easily applied to your own arguments. Your motte A is that a digital image is a collection of pixels, and that any context, history or provenance is something that people bring with them. Your bailey B is that context, history and and provenance should have no bearing on whether people feel excluded or on administrative decisions surrounding images. A does not imply B, yet you repeatedly use A to defend B.
> You just dip into this account t o be nasty without anyone knowing who you are
Can you give an example of where I have been nasty? I literally wished good things for you in the last post and have done my best at every point in this conversation to turn the other cheek. In contrast, you have - insulted my "crummy upbringing" - called into question the "moral issues" of my family - have argued I am "running around traumatizing people" - characterized my statements as "narcissistic projection" repeatedly
>You asked me to look into what drove me to write what I did, and I did just that and you overlooked it >Honestly I care so little I literally passed out, but I did inquire into myself as to why I even bothered writing back to you. And you know what came up? I feel bad for the people in your life who have to put up with you. And then I realized you're not hiding behind your keyboard when you talk to them, using an obvious troll account on HN like you are, just look at your post history.
Your conclusion is that what drove you to write these posts is that you feel bad for people who have to put up with me. I can only comment that there is a failure of self-reflection occurring.
>And obvious troll is obvious. A 9 year old account with 17 karma
This is my main, I'm just not very active on HN. People do have a range of behaviors, and even though you appear very active on this alt account of yours, I am not similarly active or engaged with this site despite this being my main. Very often, I am happy to learn from the interesting conversations without feeling a need to comment on my own. Your assumption that the account on which I am posting is an alt is a projection of your own desire to separate your views on the "chillingeffect"[1] from your main. In general, it feels as though you have trouble acknowledging that people have habits and views which diverge from your own, and in this case have reacted with anger rather than intellectual curiosity.
>I'm not required to be kind to everyone all the time and certainly not to an HN troll. Delete your account and get real.
I really truly believe that you would benefit from discussing this thread with someone you trust in a safe space. That could be a domestic partner, a family member, a councilor, a priest, or a co-worker. Just someone whose judgement and morals you admire. Someone who will understand that one interaction does not define the entirety of your identity and who believes in your innate goodness. Give it some time to cool off and come down from the feelings of anger that you are expressing here, and be the person you want to see in the world.
It may be helpful for you to consider this excerpt from Psychology Today:
|The precursor of anger in mammals is a perception of vulnerability plus threat. The more vulnerable people and animals feel, the more threat is perceived. The function of anger is to protect vulnerability and neutralize threat.
|In humans, the threat is almost always to the ego (how we want to think of ourselves and have others think of us). Anger neutralizes ego-threat by devaluing, demeaning, or undermining the confidence of the person perceived to be threatening.
[1] https://www.mtsu.edu/first-amendment/article/897/chilling-ef...
You may be a kind and intellectually considered person in other aspects of your life, but this is not an example. May kindness find you, may people give you grace, and may you be a source of warmth to others.
I have a viewpoint and I am expressing it, in accordance with free speech. I'm sorry that my viewpoint make you "feel bad". I'm sorry that you had such a strong reaction to this image of the author, published in their own expression of free speech. Everyone could use a bit of self-examination, and I would suggest that some examination of why this article has led to such an emotional response in yourself is similarly worthwhile. Your personal attacks on my "narcissistic projection", "crummy upbringing" and being "wrapped around the axle" only serve to heighten the emotional stakes of the argument and are not a discussion of the relative merits of the issue.
I have no qualms about sex work and find stigmatization of such work distasteful. Nonetheless, a society in which women are being policed for their wardrobe in a professional settings and also confronted with images sourced from erotica in a professional setting is hypocritical. Many women are saying that this image is unwelcoming and that they feel it is exclusionary. Playboy is exclusionary and explicitly stated in its first issue that it was not for women, and it's very hard to argue that the Playboy centerfolds were not intended in large part to cater to a male fantasy. They are saying that an action can be taken which does no technical harm to the merits of JPEG compression technology and would have appreciable gains from an inclusionary standpoint.
Again, I'm sorry that them expressing their viewpoint is so distasteful to you and has evoked this response in you.
if the participants are (at least) minimally knowledgeable
Nevertheless, HN has clear guidelines about comments. Of particular relevance is
> Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.
More generally, HN can be a great place when people take the time to consider their response and do a little digging. Discussions are more meaningful and fruitful if the the participates are minimally knowledgeable on the topic being discussed. This goes for any topic. Imagine if the reply to a machine learning post was asking "what if this picture-sorting AI became sentient and starting destroying the world? How can we know the scientists working on this can outsmart the AI? How can we be sure that they are being kind to our future overlords?"? They may feel that their view is completely valid, reasonable and informed because of the wealth of Sci-Fi literature that they have read, despite never having attempted to grapple with the actual material being submitted to HN.
In this hypothetical situation where straight men had historically struggled to break into the ranks of successful computer scientists, it's likely that gay men would similarly feel that if they were to complain about too much straightness in the work place it would be a hate crime.