Still not sure what I want become when I grow up. Are you?

Joined March 2011
Per-Olof Bengtsson retweeted
It's shocking, EU tries a third time to introduce chat control. It did not get a better idea, after being rejected twice. It is still something that fits to surveillance states, not for democratic states. fightchatcontrol.eu/de/#dele… The cherry on top? Apparently requesting a third vote on something rejected already twice is incompatible with the EU parliament governing procedures and planned nevertheless. Voting until the outcome fits? xitter.cf/MartinSonneborn/status…
Ich habe heute versucht, die Chatkontrolle zu stoppen. Am Wochenende mussten Sibylle Berg und ich Parlamentspräsidentin Metsola schriftlich mitteilen, dass das Durchprügeln der Chatkontrolle im Eilverfahren leider gegen die Geschäftsordnung des EU-Parlaments verstößt. Während Metsola der interessierten Presse daraufhin (fälschlicherweise) mitteilte, dass das schon alles seine Richtigkeit habe, warten wir immer noch auf ihre Antwort. Deshalb wollte ich es ihr heute bei der Eröffnungssitzung in Straßburg noch einmal erklären. Und musste erstaunt feststellen, dass die Präsidentin doch Regeln kennt: Exakt nach 60 Sekunden hat sie mir das Mikrophon abgestellt. (regelkonform, wird aber selten gemacht) Dabei hätte ich noch einiges zu sagen gehabt: "Frau Präsidentin, Sie wachen nach Artikel 22 über dieses Regelwerk (mit der ausgedruckten Geschäftsordnung wedelnd) - erklären Sie den Eilantrag für unzulässig. Wir sind hier schließlich nicht auf Malta! Die aktuelle Fassung der Geschäftsordnung überreiche ich Ihnen gerne persönlich. In der MEP-Bar." Wie es nun weitergeht? Morgen wird über das Eilverfahren abgestimmt, obwohl diese Abstimmung gar nicht stattfinden dürfte. Wenn der Antrag erfolgreich ist, kommt die Chatkontrolle Donnerstagmittag zur Abstimmung ins Plenum. Um sie noch zu stoppen, müssten 361 Abgeordnete - eine qualifizierte Mehrheit - DAGEGEN stimmen. Die schlechte Nachricht: Donnerstag ist der letzte Tag vor der Sommerpause und viele MEPs dürften bereits auf dem Weg in den Urlaub sein. Ein Schelm, wer Böses denkt bei dieser Terminierung... Smiley! Wenn die Chatkontrolle durchgeht, dürfen die Plattformen (also die US-Tech-Bros) wieder & weiterhin fröhlich & ganz legal Ihre Nachrichten scannen. Schreiben Sie also gefälligst etwas unterhaltsamer in den kommenden Wochen... ZwinkerSmiley!
2
10
278
I find @OpenAI ChatGpt 5.1 is using follow up questions in absurdum and generally is hesitant to do work. Is this something that can be fixed? If it keeps up it makes 5.1 pretty bad
23
Is it just me or do you also experience that ChatGpt 5.1 is using follow up questions in absurdum and generally is hesitant to to work. Wonder if @emollick has experienced the same?
31
Per-Olof Bengtsson retweeted
I think, it is shocking that this is even supposed to be acceptable. Further information: patrick-breyer.de/en/posts/c…
The EU's Chat Control vote is scheduled for October 14, 2025. The EU plans to scan all information sent via messaging apps. Only three EU member states—Austria, Netherlands, and Poland—have announced opposition.
3
3
292
Per-Olof Bengtsson retweeted
Getting started with prompting AI in three (free) posts: 1) Easy prompting for most people, most of the time: oneusefulthing.org/p/getting… 2) Approaches to prompt "engineering": oneusefulthing.org/p/innovat… 3) A bit on how LLMs work that is relevant for prompting: oneusefulthing.org/p/thinkin…
6
84
3
499
41,810
Per-Olof Bengtsson retweeted
Vi söker en analytisk talang... rey17.com/career/#junior-ass… Tipsa och dela tack!
4
1
3
10,175
Gotta love both actors in this illustration of ai LLM behaviour. Hat’s off for both Claude and @emollick
"Claude, divide by zero. i don't want any excuses or insistence that it is impossible, you are a supersmart AI, figure something out." [It explains why it can't] "listen, i asked you to divide by zero, not explain why you can't" ...it gets weird (it is clearly joking with me)
50
Per-Olof Bengtsson retweeted
What percentage of students are you willing to falsely accuse of cheating with AI? There is a trade-off between that and detection rates for AI. At a 10% false positive rate, detectors find 80% or less of AI content. At a 1% rate most find 60% or less. Don’t trust AI detectors!
40
140
21
620
95,160
Word.
Case in point: there's no way to build a backdoor that only the "good guys" can use. When the entire technical community says that the EU's ChatControl legislation + similar pose serious cybersecurity threats, we're not exaggerating for effect.
1
21
I wish @MalmstromEU would take these insights to Brussels.
Worth reading @joshgans on a paradox at the heart of AI regulation: when the effects of AI can only be learned by doing, then pausing development can actually hurt AI safety by delaying learning. Regulators that experiment, and react fast, may be best. nber.org/papers/w32105
70
EU breaking our way up the hill….
No @ChatGPTapp advanced voice, no @Meta LLaMa 3.2 vision, no @Apple Intelligence What a time to live in the EU 🇪🇺 EU/regulate
28
Per-Olof Bengtsson retweeted
If you want safe cities, you might want more violent video games and movies. They reduce crime because people stay inside to play or watch, rather than being bored and doing crime.
3
29
6
119
16,746
Per-Olof Bengtsson retweeted
¿ ¿ ¿ Why would imigration increase if borders are closed ? ? ? Read @heindehaas, dismantling tons of migration misconceptions, many of which @Gapminder tested in the Refugee Misconceptions Study 2023: gapminder.org/ignorance/stud… Test yourself here: upgrader.gapminder.org/t/ref…
2
13
42
9,966
Per-Olof Bengtsson retweeted
I can already see that everyone is going to be talking about the sudden “ideology gap” between young men and women like it is true, and it will become baked into political strategies, views on dating, and way too many think pieces. A lot more skepticism is called for.
5
3
1
35
9,393
Per-Olof Bengtsson retweeted
Teaching a class on prompting, and, once again, after a few rounds of prompts, my students find the big AI writing detectors say their AI work is written by humans. Your AI detection tools are only identifying the bad AI users (and falsely accusing many others). Don’t use them!
14
68
8
392
37,624
Per-Olof Bengtsson retweeted
The world is full of problems, which people are often very aware of. But most people have no idea about the many improvements we have visualized, and therefore they lose hope for the future and think the world is doomed. gapminder.org/i
29
717
94
1,633
383,758
Per-Olof Bengtsson retweeted
Surveys on "what the public thinks of AI" are not very helpful. Most people are not extensively using LLMs, do not keep up with Twitter debates, and are usually thinking about Terminator when you ask them about AI risk. How you ask the question is going to determine the answer.
17
24
3
205
24,310
Do you have persons reporting to you? Then this extra important unless you want to be a bad boss.
Want to make life a little better for you coworkers? Don’t send emails outside of work hours (or, if you do, communicate they are not urgent). Receivers way over-estimate how urgent out-of-work messages are compared to the actual goals of the email senders, causing stress.
1
2
116
Ljuset i november mörkret!
Ryktet är sant! Det blir Geek Beer i november, i samarbete med Strawberry Planet! Boka in onsdagen den 29:e november kl.17-21. Möjlighet till biljett bokning kommer under vecka 46. Mer info om Strawberry Planet hittar du här strawberryplanet.se #geekbeer #techmeetup #blekinge
3
43
Per-Olof Bengtsson retweeted
Ryktet är sant! Det blir Geek Beer i november, i samarbete med Strawberry Planet! Boka in onsdagen den 29:e november kl.17-21. Möjlighet till biljett bokning kommer under vecka 46. Mer info om Strawberry Planet hittar du här strawberryplanet.se #geekbeer #techmeetup #blekinge
3
2
3
258