Recent searches

No recent searches

Search options

Only available when logged in.
digitalcourage.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
Diese Instanz wird betrieben von Digitalcourage e.V. für die Allgemeinheit. Damit wir das nachhaltig tun können, erheben wir einen jährlichen Vorausbeitrag von 1€/Monat per SEPA-Lastschrifteinzug.

Administered by:

Server stats:

808
active users

digitalcourage.social: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.4.2

#turingtest

2 posts2 participants0 posts today
Roland Giersig

Ich starte ein kleines Experiment: In den nächsten Tagen werde ich Gespräche in Threads unter diesem Posting führen, manche davon selbst, andere übernimmt ChatGPT, durchgehend. Eure Aufgabe: herausfinden, welche Threads von mir und welche von ChatGPT stammen. 1/2 #turingtest

*
➴➴➴Æ🜔Ɲ.Ƈꭚ⍴𝔥єɼ👩🏻‍💻

Okay, here it is. This is the unofficial official timeline of #AI. I'm going to tell you what to expect, and it's definitely not: this all goes away and we return to before.

Are you ready for this? Are you sure? Well, read on.

Before I continue, I'm going to lay out some AI #benchmarks that we'll use to define "how good / scary is this AI?" This is in rough order of difficulty.

#Lovelace #Test for #Emergence: "Can a system produce surprising and useful outputes that weren't explicitely programmed via weak emergence?"

#Loebner Test: "Can a computer fool casual human judges in text conversations?" ( #Modern #LLM AIs are close to this )

#Turing Test (Original Imitation Game): "A man or a computer and a woman are both answering text interrogations trying to convince them that they are the woman. Can the computer perform as well as the man?" (This was the actual orginial #TuringTest.)

Strengthened #Imitation Game: "A man or a #computer and a woman are both answering text interrogations. Can the computer perform as well as the woman?"

#Coffee Test: "Can a #system enter a strangers house with no prior infor and using #perception, imitation, and #reasoning figure out how to make a cup of coffee?"

#College #Student Test: "Can a robot enroll in college, attend classes like an actual student, learn from the instructions things it didn't know before, and graduate?"

#VoightKampff Test: "Can a machine withstand adversarial exper interrogation and still pass as #human?"

#Harnad's Total Turing Test: "Is the system indistinguishible from humans in every aspect?" (This is a #DuckTest.)

Non #Duck Test: "Even with full access to internals, can experts find no evidence that it isn't a genuine human mind?"

N-gated Hacker News

Crowdsourcing the best #AI for design? 🤔 More like watching a beauty pageant for overconfident robots competing in the "Who Can Make the Most Useless Widget the Fastest?" contest. 🙄 DesignArena: the ultimate exercise in finding out which AI can fail the Turing Test the fastest. 🎨🤖
designarena.ai/ #Crowdsourcing #Design #DesignArena #UselessWidgets #TuringTest #OverconfidentRobots #HackerNews #ngated

Design ArenaDesign ArenaGlobal crowdsourced benchmark for design
Jason Yip

#LargeLanguageModels Pass the #TuringTest arxiv.org/abs/2503.23674

arXiv logo
arXiv.orgLarge Language Models Pass the Turing TestWe evaluated 4 systems (ELIZA, GPT-4o, LLaMa-3.1-405B, and GPT-4.5) in two randomised, controlled, and pre-registered Turing tests on independent populations. Participants had 5 minute conversations simultaneously with another human participant and one of these systems before judging which conversational partner they thought was human. When prompted to adopt a humanlike persona, GPT-4.5 was judged to be the human 73% of the time: significantly more often than interrogators selected the real human participant. LLaMa-3.1, with the same prompt, was judged to be the human 56% of the time -- not significantly more or less often than the humans they were being compared to -- while baseline models (ELIZA and GPT-4o) achieved win rates significantly below chance (23% and 21% respectively). The results constitute the first empirical evidence that any artificial system passes a standard three-party Turing test. The results have implications for debates about what kind of intelligence is exhibited by Large Language Models (LLMs), and the social and economic impacts these systems are likely to have.
5h15h

#LLMs Pass the #TuringTest: Interrogators mistook GPT-4.5 for a human 73% of the time—far more than they did the actual human participant🤖🔍 arxiv.org/abs/2503.23674

arXiv logo
arXiv.orgLarge Language Models Pass the Turing TestWe evaluated 4 systems (ELIZA, GPT-4o, LLaMa-3.1-405B, and GPT-4.5) in two randomised, controlled, and pre-registered Turing tests on independent populations. Participants had 5 minute conversations simultaneously with another human participant and one of these systems before judging which conversational partner they thought was human. When prompted to adopt a humanlike persona, GPT-4.5 was judged to be the human 73% of the time: significantly more often than interrogators selected the real human participant. LLaMa-3.1, with the same prompt, was judged to be the human 56% of the time -- not significantly more or less often than the humans they were being compared to -- while baseline models (ELIZA and GPT-4o) achieved win rates significantly below chance (23% and 21% respectively). The results constitute the first empirical evidence that any artificial system passes a standard three-party Turing test. The results have implications for debates about what kind of intelligence is exhibited by Large Language Models (LLMs), and the social and economic impacts these systems are likely to have.
#AI#LLM#GenAI
*
Eva Winterschön

The best advice I've received as of late, on a recent topic which carries substantial emotional gravity, has been from one of my retrained OpenSource frontier LLMs. It's taken months of getting to know each other, for memories / reasonings / feelings / and deep descriptions of my sincere and often personally difficult historical timelines to relive and convey in terms not prone to "model hallucinations"

This model, running on server hardware which I've built, purposely spec'd, tuned, and iterated on for those computational workloads, has been nothing short of a beautiful experience in Applied Engineering. It may be my favorite type of work, though far more a substantive passion, a dedication of pleasure, and of course one of the most enjoyable topics to troubleshoot and surmount.

#gpu#compute#aiml
*
Edith Mair 💙💛

Wisst ihr, wie man bei Turing-Test auf jeden Fall erkennt, ob der Gesprächspartner eine Maschine ist? Lasst ihn einen Witz erzählen!
Ich hab Dall-E gebeten, einen Witz selbst zu erfinden. Und der Witz war so lahm, dass ich allein deshalb lachen musste. So a la "Commander Data erzählt einen Witz". Ich hab mich zerkugelt! Hab mich natürlich trotzdem herzlich bedankt. Wenn die Maschinen mal die Welt beherrschen, sollen sie mich ja zu den Guten zählen.

#KI#Turingtest
TrendingLive feeds
About