Beep boop - this is a robot. A new show has been posted to TWiT…
What are your thoughts about today’s show? We’d love to hear from you!
1 Like
Regaerding Twitter and it having laid off all (or most) of its staff in the area of moderation and compliance, they, along with Telegram, were last week officially charged with ignoring a majority of take-down requests from users under the German NetzDG (Netzdurchsetzungsgesetzt, which covers things like hate speech, fake news, racial hatred, incitement to violence). If they do not comply within the next couple of weeks, it will cost them $50M.
The problem is, the fine is so small, it is probably less than re-instating a full team to cover such things.
It is interesting, I filed my first ever takedown notice with Twitter yesterday. I’ll be interested to see if it goes anywhere - it was a disturbing video of a car running a cyclist of the road and the passenger, at the last second, shoving the rider into a bunch of trees, with the comment “haven’t we all wanted to do this?” Given the rider was probably seriously hurt, if not critically injured or killed, some of the comments coming off of it “he deserved it” because “he probably shouted abuse at the driver, before they started filming” don’t make it any better.
Normally, I’d block the account, but having read a couple of hours earlier about Twitter’s non-compliance with NetzDG, I decided to actually report this incitement video and see what Twitter’s reaction was.
2 Likes
Just read an interesting article in c’t. Scientists at Stanford University performed a study of AI research.
China is way ahead of the rest of the world, when it comes to AI research.
MIT appears at spot ten on AI research institutes, the first 9 places are all taken by Chinese institutions - the authors of the study looked at the published papers from 2010 to 2021.
Interestingly, this all changed in 2022, when the index of the 35 most influencial AI models suddenly switched from academia to commercial buisinesses (as we’ve seen in the press since the middle of last year), with 32 of the most important/influencial models suddenly coming from business and only 3 from academia.
The top 10 list is:
- Chinese Academy of Sciences
- Tsinghua University
- University of Chinese Academy of Sciences
- Shanghai Jiao Ton University
- Zhejiang University
- Harbin Institute of Technology
- Beihang University
- University of Electric Science and Technology
- Peking University
- Massechusetts Institute of Technology
MIT peaked at position 7 in 2020, having climbed to 8 the year before, before falling back to position 10 in 2021.
Despite the sudden commercialisation of AI in 2022, investment sunk to $91.1 billion. a drop of 26.7% over 2021. But looking back over the last 10 years, investment in 2022 was still 18 times as high as in 2013.
In an IPSOS survey, 78% of Chinese felt like AI had more benefits than disadvantages, the best of any of the 81 countries surveyed. The USA brought up the rear, with only 35% of people seeing the benfits outweighing the disadvantages.
Analysis of the AIAAIC database also showed fast growth in the misuse of AI. According to the figures, the number of registered cases of incidents of AI being misused rose by a factor of 26 between 2012 and 2022.
Ecologically, AI doesn’t do very well either. The training set for the AI speech model “Bloom” uses 25 times the amount of CO2 as a person flying from New York to San Francisco.
They also state that existing benchmarks for AI are no longer relevant, moving forward and that for future reports, they will need to move over to Google’s BIG-Bench and Stanford’s Holistic Evaluation of Language Models (HELM) benchmarks.
The full report is available as a PDF from Stanford:
And those interested in the data behind the report, they have released that as a Google Suite table:
https://drive.google.com/drive/folders/1ma9WZJzKreS8f2It1rMy_KkkbX6XwDOK
The only services that were invite only anf people are still talking about seem to me to be Gmail and Spotify. Does anyone still use Clubhouse? So, maybe, that isn’t a model that Bluesky ought to have followed
The reason I left was I had a message from a blue checked Elon musk with a picture of 3 kids, 2 apples and a knife with a comment saying with one cut of the knife share the apples equally, there then followed an image showing a bloody knife and one of the kids dead!!!
Every senior student doing computing studies in China has to do several modules on AI, while here in the UK we play with Microbits turning LEDs on and off!!!
I want to comment about @JeffJarvis statement that AI and Bing AI in particular is rubbish, I would also make mention of him rubbishing Wikipedia.
There is a phrase among the HiFi fraternity ‘garbage in, garbage out’ in other words the end result is only as good as it source material and the internet is full of cats, porn and opinions.
It is a world where influencers, most of whom, have no knowledge of the tat they prattle on about, not forgetting how easy it is to pay for top billing on Google or Amazon.
The internet is not reality, neither are the TV shows that show everyone uses Apple computers it is simply a paid, skewed view of the world, so what would you expect an LLM trained on it to give you?
I have very successfully used Bing AI to plan a getaway to Amsterdam, and am using it to work on a trip to Copenhagen. We are also using to plan an end of year getaway to the west coast and I have to say, it seems pretty accurate - it even told me about Petaluma!!
The thing most of the critics don’t seem to pick up on is the fact that you can use Bing as a traditional search engine and then if you want the bigger picture click on ‘chat’ to more interactive.
1 Like
That is only because of the film set in Petaluma 
(Hmm, I watched it a few months back, but not showing up on IMDB. A girl having problems with her step father runs away to her real father, a down-and-out mechanic who used to race. He lives in Petaluma and she turns him around & starts racing herself at the Petaluma Raceway…)
1 Like
It rambles on about a seedbank in Petaluma as well
1 Like
Curious to see what happens for you. I gotta tell ya, BEFORE the Musk-administration, i had issues with take-down requests and so forth. Nasty, nasty things have been said about me. I have thick skin, but some things went too far. It was never an immediate take down. It was a “process” that could take a few days if I’m lucky.
3 Likes
Just finished this weeks show, and one thing sticks out. Last week, Leo talked about how Elon manipulates the media into giving him free publicity. That segment went on for forty minutes. This weeks show covered almost an hour of Twitter and Twitter adjacent social media. Rule 1 of internet trolls is dont feed the troll and they will eventually go away. At this point, TWIT has been on an “everything we hate about Twitter” holy war for months, combined with a push to get people on mastodon. Please, please, please, just stop talking about Twitter and the problem will sort itself out.
To quote one of your show titles from earlier this year, “Stop trying to make mastodon happen.”
I thought the second half of the show was thought provoking and covered some good material.
Final addition - Briana discussed the question of the US as a democracy. If you do a word search of the US Constitution, it never mentions democratic or democracy, but it does say “The United States shall guarantee to every State in this Union a Republican Form of Government,” so the correct answer is Republic, or as Ben Franklin said, “A Republic, if you can keep it.”
"Full Text of the U.S. Constitution | Constitution Center
3 Likes
Not true. I personally still utilize Twitter as a “broadcaster”. As does our TWiT marketing squad. I don’t use it much for “social,” but it has it’s place and I’ve been up front about that on any show it’s brought up and I’m on air.
3 Likes
I laugh to myself when i mention mastodon to other folks outside of this tech world we live in. As soon as I mention “federated” or “pick a an instance,” their eyes glaze over. Lol! Even if i simplify it in explanation, it always comes back to “well who’s on it?”
3 Likes
Under German law, they have to take down violence, incitement or hate speech within 24 hours of being notified, if it isn’t clear or is a lesser offence, they have 72 hours to comply.
That was the problem in Germany, they had failed to take down ove 80% of valid claims, which is why they are facing a fine.
2 Likes
Well, that response wasn’t within 24 hours, but within 72 hours…
I just got a response from Twitter. The account I reported has been banned for breaking Twitter’s rules regarding incitement to violence…
Edit: Regarding AI, I tried using Bing AI to find the text for a German song sung by kids who go collecting on 10th November (Mattemänken) - it is done here in the region instead of Trick or Treat.
Bing replied that the song is as follows:
Mattemännken goue man Goue man, goue man Mattemännken goue man Goue man, goue man Mattemännken goue man Goue man, goue man
Wat häst du in de Tasse? In de Tasse, in de Tasse Wat häst du in de Tasse? In de Tasse, in de Tasse Wat häst du in de Tasse? In de Tasse, in de Tasse
Kaffee un Kauken Kauken un Kaffee Kaffee un Kauken Kauken un Kaffee Kaffee un Kauken Kauken un Kaffee
Dat will wi nich hebben Nich hebben, nich haben Dat will wi nich haben Nich haben, nich haben Dat will wi nich haben Nich haben, nich haben
Wat willt du denn hebben? Denn haben, denn haben Wat willt du denn haben? Denn haben, denn haben Wat willt du denn haben? Denn haben, denn haben
Äppel un Birnen Birnen un Äppel Äppel un Birnen Birnen un Äppel Äppel un Birnen Birnen un Äppel
Dat kanns du wol kriegen Wol kriegen, wol kriegen Dat kanns du wol kriegen Wol kriegen, wol kriegen Dat kanns du wol kriegen Wol kriegen, wol kriegen
Nu giff mi dat Äppelken Dat Äppelken, dat Äppelken Nu giff mi dat Äppelken Dat Äppelken, dat Äppelken Nu giff mi dat Äppelken Dat Äppelken, dat Äppelken
Nu giff mi dat Birnchen ook Dat Birnchen ook, dat Birnchen ook Nu giff mi dat Birnchen ook Dat Birnchen ook, dat Birnchen ook Nu giff mi dat Birnchen ook Dat Birnchen ook, dat Birnchen ook
Nu dank wi di so schön dafür So schön dafür, so schön dafür Nu dank wi di so schön dafür So schön dafür, so schön dafür Nu dank wi di so schön dafür So schön dafür, so schön dafür
Un wünsch di allns wat goot is Wat goot is, wat goot is Un wünsch di allns wat goot is Wat goot is, wat goot is Un wünsch di allns wat goot is Wat goot is, wat goot is
Un wenn wi wedder kommen Wedder kommen, wedder kommen Un wenn wi wedder kommen Wedder kommen, wedder kommen Un wenn wi wedder kommen Wedder kommen, wedder kommen
Dann willt wi mehr hebben Mehr hebben, mehr hebben Dann willt wi mehr hebben Mehr haben, mehr haben Dann willt wi mehr haben Mehr haben, mehr haben
It had tried to transcribe a video children singing the song, but it seems that it couldn’t differentiate between the voices and bring out a single voice or a single text, so it interpreted every word sung by every child at the same time. It probably doesn’t help that the song was in Plattdeutsch (Flat German) which is a dialect of North West Germany, where I live.
I managed to eventually track down a version of the text in a newspaper article and taught Bing AI the correct words. It thanked me and eventually came up with some additional information.
The correct text is as follows:
Mattemännken gohe Man,
ob du uns was geben kann,
Appel odder Bireren,
Nörte sind vögierwen,
aule Statd - nigge Stagt,
leewe Jungfrau giff us wat,
laut us nich tou lange staun,
wi mört noch’n Hüsken wiedergaun,
bet nau Briärmen,
do wült wi us’n bittken wiärmen,
bet nau de nigge Stadt,
do stiärk’t se us oll tohaup innen Sack
A big difference!
What was interesting, I suggested that Bing had tried to interpret all of the voices individually and in parallel and Bing agreed and said it had had difficulty trying to separate the voices…
1 Like
Social media is so complicated for the casual user now. Mastodon is confusing to many, Twitter is becoming a ghost town, and how does one even get on Blue Sky? Unless Twitter reverts to what it was, this will drive the everyday person back to Facebook. Zuckerberg has to be enjoying all this confusion.
See, all you had to do was talk nicely to it 
2 Likes
Honestly, Mastodon being hard to understand is a feature, not a bug. Just check out Bluesky for an idea of how “easy to use” can fail. And that’s still invite-only!
And, just to go on record, if I were king of TWiT I would stop all company Twitter use entirely. But I’m not. I honestly feel like any use of Twitter at all is tacit support for what it’s become: $8chan
5 Likes
You can be king of TWiT for Saturday
1 Like
I have noticed my twitter feed become not abhorrent but just incredibly bizarre in the past week. Occasionally it will be the people I follow, but more often than not it’ll just be random feeds of fails and people being stupid caught on CCTV, none of with are blue accounts either.
I need to start using Masto more to post, I read twit’s local quite a bit, but need to start putting more things out there.
1 Like