Mistä muutos johtuu?
Vain muutos on pysyvää, sanotaan. Tavallaan tämä on ihan totta – maailmahan on muuttunut aina. Samaan aikaan asia on kuitenkin mahdollista sivuuttaa. Monet meille keskeiset asiat kun pysyvät muuttumattomina vuodesta toiseen. Itse asiassa monet perustavanlaatuiset inhimilliset…
Epävarmuus ei ole häiriö – se on nykyinen normi. Siksi tulevaisuutta ei rakenneta ennustamalla, vaan vahvistamalla muutoskyvykkyyttä: resilienssiä ja antifragiliteettia.
Erehdysten hyödyntäminen on muutoskyvykkyyden ydintaito
Yksi suurimmista muutoskyvykkyyden esteistä on pelko. Itse asiassa iso osa nykyihmisten huonovointisuudesta johtuu siitä, että pelon aiheita on paljon, eikä tähän auta, vaikka lukisi sata kirjaa siitä, että suurin osa peloista on oikeasti…
Kaksi tapaa toimia muutoksessa
Olin aikoinaan Salt Lake Cityssa paneelissa opetusteknologian arvostetuimpiin kuuluvassa ASU/GSV-konferenssissa. Konferenssin pääpuhujia ovat olleet muun muassa Microsoftin perustaja Bill Gates ja ex-presidentti George W. Bush. Yksi paneelikeskustelijoista oli…
Tieteellinen ennustaminen on päivä päivältä haastavampaa
Kiinalaiskirjailija Cixin Liun science fiction -eepos Kolmen kappaleen probleema käynnistyy visaisella mysteerillä. Tutkijat ja etenkin fyysikot ympäri maailmaa ovat lannistuneet tieteellisen tutkimuksen haasteiden kanssa ja osa on päätynyt…
Actually, yes, this and a fantastic set of encyclopedias
Uusi blogipostaus mustista joutsenista ja yhä kiihtyvästä ennakoimattomuudesta:
ajattelunammattilainen.fi/2025/04/22/m...
The English translation of my book The Machine Mind has been published today. What makes this particularly interesting is that, true to the book's theme, both the translation and the audiobook have been created using generative AI.
Check out the book here: www.bookbeat.com/uk/book/the-...
🐧
👀This paper finds "the first robust evidence that any system passes the original three-party Turing test"
People had a five minute, three-way conversation with another person & an AI. They picked GPT-4.5, prompted to act human, as the real person 73% of time, above chance. arxiv.org/pdf/2503.23674
This is now a beach with no elephant created with 4o.
No, I think you're actually right. I tried to modify the picture using a prompt and it created a completely new one. That's strange, though, as in my other tests today I've managed to actually generate and modify images using the new system.
What makes you think it would be Dall-E though?
Well the generation happened transformer-style (rastering layer by layer) rather than diffusion-style (removing visual noise) so I suppose it should be native? Also, at least back in the day, it labelled the pics with Dall-E when that was used, and as you can see in the screencap, there's no label.
What do you mean?
Sorry, I take that back. On second try, ChatGPT did it right too. Perhaps not as photorealistic as Gemini, but no elephant.
Gemini's latest in AI Studio actually can. ChatGPT nope.
🚨in our new paper we ran an experiment at Procter and Gamble with 776 experienced professionals solving real business problems.
We found that individuals randomly assiged to use AI did as well as teams of two people & were happier. It also erased skill silos. www.oneusefulthing.org/p/the-cybern...
Half of “prompt engineering” was actually just prompting LLMs to act like Reasoners before the labs realized that was a thing.
(Chain of thought/think step-by-step was the first powerful prompting technique that was discovered, now Reasoners do it automatically
Agreed. I actually realized that we do need a filter, but just one that filters out all the fluff intended to overwhelm so people can really see the severity of the situation.
Meanwhile, the Gaza and Canada sleights of hand get thrown in the same bucket with Musk's purge, and I bet it's not very many people who can actually tell the difference. :/
What really wearies me down is, this has been Trump's playbook since forever, and the media still don't show any sign of catching on. Or rather, since Trump generates clicks, they're so heavily incentivized against it it's wishful thinking they'd be more picky.
Good point. Having said that, the fact that the media don't differentiate the smoke and mirrors from the substance really plays into the pocket of them.
I really think the media should cut the T+M news to a third or so & sift out the fluff so we'd get a good picture of what's really going on.
Could somebody code a Trump and Musk filter for news sites and social media? It seems nigh impossible to look into any news these days without being bombarded with the antics of the dynamic duo.
So they're actually *trying* to crash the US economy? The only two precedents I can think of are Germany in 1938 (to rile people up for a war) and USSR in 1990 (to transfer funds from middle class to oligarchs).
Weird stuff. And the Americans really voted for this?
Okay, I take that back. Just got o3 to trip on the good old “farmer, dog, sheep and book” trick.
So still not quite there yet I guess. :P
Any other tricks to try to throw at it? OpenAI’s models since o1 haven’t fallen for Jumbo the Swimming Elephant either, and so far everything I’ve thrown at o3 seems to check out.
Testing o3-mini-high, and it seems to be the first model to consistently win the “get to 21 first” game where each participant chooses a number 1–7. There is a single win condition if the beginner chooses correctly, and previous models couldn’t find it. This one can.
@benjaminjriley.bsky.social
The US Copyright office has ruled that AI/human combined work can be copyrighted as long as a human is adding, changing or selecting elements. Prompts alone do not usually produce copyrighted work. Everything is case-by-case, but the report is clear and thoughtful. copyright.gov/ai/Copyright...
