If you’re an experienced aerospace engineer who is very knowledgeable about both rockets generally and Starship, and you’d be willing to answer a few questions for me via email, can you let me know?? Feel free to PM. (Looking for a reputable source who actually works in the field and ALSO keeps up with Starship forums, not just the latter. I know plenty of fans, would like actual engineering knowledge).
“But, but … humans hallucinate all the time! How dare you set the bar for AI so much higher!”
Yes, we all remember the many stories written by human journalists that accused Woodward and Bernstein of breaking into the Watergate Hotel.
https://www.theregister.com/2024/08/26/microsoft_bing_copilot_ai_halluciation/
The Internet Archive and its 916 billion saved webpages are back online
Wayback Machine back in read-only mode after DDoS, may need further maintenance.
The weather here in Heidelberg looks as unpromising as yesterday for viewing Comet C/2023 A3 Tsuchinshan-ATLAS this evening, but there's always this nice video showing the view that the ESA/NASA SOHO mission had when the comet rounded the Sun between 7 & 10 October ☄️
Nice processing by my friend @stim3on & happy to see that he got "in movie" credit as appropriate 🙂👍
Fingers crossed for opportunities to see it myself later in the week 🤞
Original link: https://www.esa.int/ESA_Multimedia/Videos/2024/10/Comet_C_2023_A3_brightens_SOHO_s_week
#TBT to October 1604, when sky watchers spotted a new bright dot among the stars. This supernova was named for German astronomer Johannes Kepler, who studied it extensively. Our telescopes still observe its slowly expanding remnant!
Learn more: https://go.nasa.gov/4dFNVZF
Our new paper (led by newly minted graduate, Florian Lalande) on using machine learning to impute unknown exoplanet properties has just dropped!
Let's talk deets...
We're edging towards a stunning 6,000 exoplanet discoveries. It's a huge database that should be outright PLUNDERED for trends that can reveal how planets form and evolve. However, use is throttled by the fact that database has so many missing properties it's got more holes than a sponge. (cont...)
Thread on international conferences, remote attending etc...
So for the first time ever I was invited in the SOC (=Scientific Organising Committee) of a conference abroad.
Before joining, I made clear that a prerequisite for me was that a minimum amount of remote attendees/speakers should be allowed: we cannot go fully back to pre-covid.
I was reassured and so was happy to serve in the past months by setting up the program, deciding speakers etc.
Sadly...
Your 80-year old reminder of the ethical decision you make if you choose to dissociate your fantasies of space exploration & love of rockets from the politics of the people building & launching them.
Credit: Deutsches Bundesarchiv (German Federal Archive), via Wikipedia CC BY-SA 3.0
https://commons.wikimedia.org/wiki/File:Bundesarchiv_Bild_141-1879,_Rakete_V2_nach_Start_cropped.jpg
So grateful in an era where so much online advice comes from generative AI without any grounding in evidence, that someone has actually made a robot to test the best phone cables.
It may seem like a small thing, but it really matters.
Mosaic Netscape 0.9 was released 30 years ago today.
According to my notes, it went live shortly after midnight on Oct 13, 1994. We sat in the conference room in the dark and listened to different sound effects fired for each different platform that was downloaded. At some point late that night I wandered off and wrote the first version of the page that loaded when you pressed the "What's Cool" button in the toolbar. (A couple days later, Jim Clark would go ballistic...
Ever wonder when your system last updated a package when using a Debian or Ubuntu Linux? Let's check Firefox, for example. Type this:
```
grep -B4 firefox /var/log/apt/history.log
```
See https://www.cyberciti.biz/faq/debian-ubuntu-linux-find-package-installed-updated-date/ for more info.
In the past few weeks, Apple canceled the following:
a) A multi-billion dollar investment in OpenAI
b) Apple officially cancels autonomous driving permit (Apple already scrapped its self-driving electric car project after ten years of research)
c) Apple engineers confirmed that AIs don't understand reasoning and stuff like that.
While the whole tech industry is going AI-first, Apple is more cautious and isn't buying AI buzz. What do you think this indicates?
Someone in misinformation research really needs to follow-up with all the biologists and other researchers (so many grad student open science folks!!) who worked to keep floods of misinformation about covid off preprint servers as best they could. It's an invisible heroic battle that hasn't been talked about enough and I've only heard about it from personal conversations. Way more interested in this than Wikipedia editing or whatever.
I don't know what's more astonishing: the new typesetting language José Reyes explains here, or the depth and quality of the explanation.
This has been my personal experience with MS Copilot: the (small) improvements in coding speed are not worth the vigilance with which you have to review the generated code.
https://open.substack.com/pub/garymarcus/p/sorry-genai-is-not-going-to-10x-computer?r=dz8ws&utm_medium=ios
NPR: All the available evidence suggests that a miniscule number of noncitizens vote illegally in federal elections and not in numbers that would sway the outcome of any race.#news #NPR https://www.npr.org/2024/10/12/nx-s1-5147789/voting-election-2024-noncitizen-fact-check-trump
I've having the most frustrating time trying to set up a Debian repository.
I have a signing key, I have a signed InRelease file which points to Package.gz files, which point to my .deb packages.
When I run `apt update`, the repo updates successfully and the signature verifies. But when I run `apt search [packagename]`, it doesn't see my package.
I'm not even sure what to search for since there are no error messages. Anyone else hit this problem with Debian repos?
Convolutional neural networks—the machine-learning systems routinely used for image recognition—are inherently limited, according to a new study by @lamaral and colleagues. #ConvNets fail on data above a certain threshold of complexity. They lean too heavily on shortcuts—spurious correlations that don’t generalize—and on localized features such as texture, to the neglect of the overall scene. h/t @manlius https://doi.org/10.1063/5.0213905 #AI
Theoretical physicist by training (PhD in quantum open systems/quantum information), University lecturer for a bit, and currently paying the bills as an engineer working in optical communication (implementation) and quantum communication (concepts), though still pursuing a little science on the side. I'm interested in physics and math, of course, but I enjoy learning about really any area of science, philosophy, and many other academic areas as well. My biggest other interest is hiking and generally being out in nature.