Nick @internic@social.cleverlibre.org

@hacks4pancakes Sounds sort of like an old Vorlon saying.

@RnDanger @boomer001 @arstechnica I was thinking perhaps it could be explained nontechnically by telling people that when they ask a #LLM a question X it isn't answering that question; rather it is answering the question "what might a typical answer to a question similar to X look like?"

@arstechnica '... he asked the AI tool whether Varghese is a real case. ChatGPT answered that it "is a real case" ... When asked if the other cases provided by ChatGPT are fake, it answered, "No, the other cases I provided are real and can be found in reputable legal databases such as LexisNexis and Westlaw." '

If accurate, to me this illustrates the core of the issue with many people's use of so-called #AI tools: this lawyer bought the hype that these #LLM are verging on human-like intelligence and could understand and answer his question, while in reality they were only constructing a statistically probable sequence of words that might be written in response to such a query, with no relationship to the facts of the matter.

I'm not sure that I agree with this, but I definitely find it interesting that Heisenberg said it.

@whknott @gutenberg_org I saw a number of people suggest adding verification to their profile without further guidance or explanation. I'd suggest to folks that if you feel the need to comment on that, then take the time to link to an explanation of how one can accomplish it.

@johncarlosbaez @lmorchard @mike @tchambers @MartinEscardo with feeds being primarily driven by chronology, I think it's just too easy to miss things if you follow more than a few people. There are potentially ways to cope with that somewhat, using lists and hashtags, but I don't think you'll ever get enough people to do that consistently to promote conversation the way one might like.

Based on past experience on social media, people are rightly apprehensive about algorithms, but perhaps they have perhaps over-corrected. It seems plausible that transparent algorithms under the control of users themselves could be a good thing. Otherwise we are effectively stuck using a very poor algorithm to decide what we see.

@lmorchard @tchambers @MartinEscardo

@lmorchard @mike @tchambers This discussion made me think of the following discussion between @MartinEscardo, @johncarlosbaez, and others about how mastodon sometimes seems to fail to surface content to those interested in seeing it (in this case specifically on category theory).

https://mathstodon.xyz/@MartinEscardo/110396681404988001

I do think this is basically due to algorithm phobia.

@ingrid I mean, it also seems weird because on a computer very capable of modeling 3D space there's presumably much less need to use a planar projection at all.

@idoubtit I was thinking recently of the relatively stringent permitting requirements for building a shed where I live and wondering if maybe that's part of the reason so many people use their garages for storage.

@dmacphee @helenczerski Yeah, I'm not really interested in quibbling over terminology either. My point was that I consider being a #STEM researcher to be a distinct skill set from communicating STEM information to the public effectively, whatever you want to call that latter role.

In my experience plenty of STEM researchers think they are capable of the latter, but they are often very much mistaken. #scicomm is based on somewhat different aptitudes and requires dedicated thought and work to come up with evidence-based methods for effective communication with different audiences.

I was just shown this PasswordPocket "offline password manager" and I'm trying to figure out what to make of it.

https://www.kickstarter.com/projects/atlancube/passwordpocket-your-personal-data-guardian/description

They say their whitepaper is forthcoming, so I wouldn't put any trust in it until knowledgeable #infosec people have reviewed that. The information in the Kickstarter seems pretty vague to me (what are the details of the E2EE they keep mentioning, what parameters are used with PBKDF2, what mode is AES256 being used in, etc.) .

I guess I'm also not convinced it is much better than using KeePass and a thumb drive, since it sounds like in general as soon as you use the backup function with a device it will have access to both the master password (in memory) and the encrypted vault. I'm definitely interested in what others think.

#security #password #passwords #passwordManager

@helenczerski If I do take your meaning to be that science communicators are not needed, I disagree.

Firstly, I think the premise that other fields like public policy and history don't have communicators is false. I believe, for example, that there are numerous people who basically make a living out of making public policy accessible to the public (e.g. the folks on the Vox podcast The Weeds).

Secondly, I think STEM fields are dissimilar from fields like history (to use your example) in three ways that increase the need for communicators:
1. STEM concepts tend to build cumulatively as some things are understood and used as the basis to explore new areas
2. Communications between STEM practitioners often use math as a fundamental component. This is essentially an entirely different language that the general public is much less comfortable with.
3. STEM often deals with situations very far from human experience, e.g. the microscopic or subatomic, galactic scale, or compounds or processes that only exist under conditions incompatible with human life. So not only are the theories unfamiliar, but the very constituents under discussion are unfamiliar (and often unintuitive in behavior).

Finally, while scientists do have to communicate their findings, it's generally to colleagues in the same sub-field or program managers for grants in that sub-field, and communicating to someone whose perspective is already almost the same as yours is the easiest case. Communicating to someone coming from a different perspective, especially engaging them if that person is not even sure they're interested, is a very different skill. In my experience, the latter skill seems to be largely uncorrelated with the ability to do research.

But I feel like you must know all of the latter very well from your own experience, so perhaps your point was something different.
#science #scicomm

@mhoye @vmstan @will@supercool.fun I have an ASUS router, and it did become unresponsive and need a hard reboot the other day, but it worked fine after that, so I didn't question too much. Now I am wondering if this is the time to try the Merlin #OSS for it.

I'm just wary at this point of using anything for which open source is not a fallback option (if for no other reason than you never know when a company will stop delivering security patches). And I'm not sure I'm ready to invest the time to setup and manage a pfSense box.

@mhoye I was going to ask if you'd considered ASUS products, since many are comparable with #FOSS firmware (as an out if things with the company get sketchy), but then I saw this:

https://mastodon.social/@arstechnica/110397243062385827

@vmstan @will@supercool.fun

Everyone pray for those Germans in your life, because we have again entered the time of the year for...Spargel madness!

@helenczerski I'm not certain which of the following you're saying:
1. Science communicators don't (presently) need to exist, because scientists can do that job.
2. Science communicators wouldn't need to exist if scientists trained on and valued communication properly.
3. You agree that a specialization in science communication is necessary, you just don't like the term "science communicator".
4. Something else entirely.

I read your post as number 2, but I'm not positive that's right.

@feditips How is "Active moderation against racism, sexism, homophobia, transphobia" determined?

@idoubtit Having taken surveys for psychology experiments in college, it was clear to me that the Myers-Briggs tests I saw were not nearly as rigorous as real psychological surveys. I was curious to learn about research on the "Big Five" personality traits, which I think should still be taken with a grain of salt but is at least much more evidence-based.
https://en.m.wikipedia.org/wiki/Big_Five_personality_traits

@vicgrinberg I've been there. I once did a week long kayak camping trip and then two weeks later hurt my lower back while ironing a shirt.