techradar (@techradar@c.im)
#TechRadar I finally got ChatGPT to stop asking 'Want me to…' at the end of every response – here’s how to do it https://techrad.ar/4tja #ChatGPT #OpenAI
Show Original Post
Lele8182 (@Lele8182@mastodon.uno)
Anche sta volta #chatgpt chi ha azzeccato , creato un piccolo programma , che legge il file in ram , e scrive su floppy o su #SD2IEC un file di testo in formato #seq . Essendo un file di testo lo si può aprire ovunque . Successivamente, chatgpt ha creato un programma che apre un file seq mantenendo il formato testo . Il passo successivo ? Selezionare il file seq aprirlo come testo e convertirlo come file prg eseguibile . Direttamente da un vero #c64 … @computer #retrocampus
Show Original Post
rpbook (@rpbook@gts.phillipsuk.org)
I'm in a #SecondWorldWar themed Discord server, where someone asked which tank had the most variations. Its an interesting subject to discuss, but somebody else asked #ChatGPT, and seemed to think they had a definitive answer: Sherman.
Until someone else pointed out that the list of Sherman variants from ChatGPT included the Crocodile. And that the Crocodile was a Churchill variant, not a Sherman. 🙄
Show Original Post
hebrewbyinbal (@hebrewbyinbal@babka.social)
ChatGPT’s English is great. Its Hebrew? Not so much 🤖
AI won’t admit “I don’t know.” It just gives wrong answers—grammatical Hebrew no Israeli would say, textbook Hebrew instead of real conversation, stripped of nuance and culture.
Worst of all, you don’t learn how to think in Hebrew. You just memorize translations you can’t use in real life.
Hebrew isn’t about copying answers—it’s about building the logic and patterns native speakers use.
ChatGPT is a tool, not a teacher. If you want real Hebrew, you need someone who knows the difference.
⬇️ Ready to learn the right way?
#chatgpt #languagelearning #hebrew
Show Original Post
tyzbit (@tyzbit@toot.now)
if you're using a language generator, review its output and do not stop reviewing until you have found at least one error, because I guarantee you it is there.
if you find the error within the first minute, rest assured there are at least 10 more errors for you to find.
#ai #llm #slop #gemini #claude #copilot #chatgpt
Show Original Post
objects (@objects@alceawis.com)
When Maggy uploads a anti #LLM video you know its gonna be good
(This video is a bit #scary
)#ChatGPT #DeepSeek #AI #SamAltman
https://www.youtube.com/watch?v=dsDcQmJHwQc
#repost •acws #acws #LarryPage #ElonMusk
Show Original Post
GripNews (@GripNews@mastodon.social)
🌘 為 ChatGPT 注入自我懷疑
➤ 透過提示工程，讓 AI 更聰明、更可靠
✤ https://justin.searls.co/posts/sprinkling-self-doubt-on-chatgpt/
本文作者分享如何透過一個特定提示（prompt），大幅提升 ChatGPT 的回應品質。這個提示旨在讓 ChatGPT 變得更具批判性、更謹慎，並在給出答案前進行自我審查。實測發現，ChatGPT 的回應開始帶有更多的不確定性表達，思考時間顯著拉長，並且會主動進行「紅隊演練」（red teaming）來驗證自身答案的正確性，整體產出更加有用。
+ 這真是個有趣的實驗！我也想試試看能不能讓 ChatGPT 變得更「人性化」一點，減少那種一本正經卻出錯的情況。
+ 「紅隊演練」這個概念很棒，讓 AI 自己找錯，實在是很聰明的做法。不過思考時間變長，不知道會不會影響即時性？
#人工智能 #ChatGPT #提示工程 #效能提升
Show Original Post
CuratedHackerNews (@CuratedHackerNews@mastodon.social)
Sprinkling Self-Doubt on ChatGPT
https://justin.searls.co/posts/sprinkling-self-doubt-on-chatgpt/
Show Original Post
h4ckernews (@h4ckernews@mastodon.social)
Sprinkling Self-Doubt on ChatGPT
https://justin.searls.co/posts/sprinkling-self-doubt-on-chatgpt/
#HackerNews #Sprinkling #Self-Doubt #on #ChatGPT #selfdoubt #ChatGPT #AI #ethics #techdiscussion #innovation
Show Original Post
DoomsdaysCW (@DoomsdaysCW@kolektiva.social)
Is #AI Making Us Stupider? This Study Certainly Thinks So
New research highlights the impact of AI on #CriticalThinking skills.
Posted August 21, 2025 | Reviewed by Michelle Quirk
Key points
- Using AI to do the thinking impacts not only quality of work but also the long-term acquisition of skills.
- AI can have a place in education, but it needs to be carefully employed.
- Employing AI is having a detrimental impact on our ability to think critically.
Excerpt: "A new study by #MIT has published initial results highlighting a likely connection between large language models (#LLMs)—colloquially grouped under the banner of AI—and a direct cognitive cost, particularly when it comes to our ability to think critically. They conducted a study using a pool of 54 participants, divided into three groups. The groups were asked to deliver an SAT-grade essay, with the first group granted access to #ChatGPT, the second to traditional #GoogleSearch, and the third with no search resources at all. Group 3 was referred to as the '#BrainOnly' group.
"Over time, each participant had to produce several essays, and each time was attached to an EEG to record brain activity across 32 regions. Consistently, ChatGPT users had the lowest brain engagement for every essay, and worryingly showed decreasing brain activity over time, as they progressed through the essay assignments. What this represents is the ChatGPT participants getting '#lazier' with each subsequent essay, with many resorting to copy-and-paste content by the end of the study. The study also recorded that this group 'consistently underperformed at #neural, #linguistic, and behavioral levels.' The findings make sense; with ChatGPT doing much of the work, it's easier on our #cognitive processes, and it is unsurprising that it takes less mental effort.
"Perhaps the more concerning aspect of the findings is the connection between #ReducedBrainActivity and impact on long-term learning and memory. As well as recording the participants' level of satisfaction, #curiosity, #creativity and #memory as part of the study—which were all high for both the brain-only and the search assisted groups (groups 2 and 3)—a later part of the study asked each group to reproduce one of their earlier essays, with the ChatGPT group asked to do so without access to any tools. There was little evidence of this group's participants remembering much of their previous essay, highlighting the very necessary link between #CognitiveEngagement and memory and, therefore, the potential impact on #LongTermLearning and gaining new skills. Conversely, and perhaps reassuringly, however, the brain-only group was then permitted to rewrite their essay with access to ChatGPT, with their efforts showing higher levels of creativity and stronger arguments, while retaining original thinking and unique language. This could present reassurance that, employed properly, AI has a place in enhancing learning without diminishing cognitive capability.
"Despite awaiting peer review and using only a small participant sample, the research team decided to release their findings in advance of peer feedback, in what they highlight as an urgent consideration of #LearningWithAI."
#DumbingUsDown #AIIsDumbingUsDown #AISucks #UseYourBrain #CriticalThinkingSkills #EvaluatingSources #UseAIWithCaution
Show Original Post
riffreporter (@riffreporter@mastodon.social)
Viele nutzen #ChatGPT als Alltagshelfer. Warum die #KI nicht auch für die #Urlaubsplanung einsetzen? Stefanie Unbehauen weiß, worauf es dabei ankommt: https://www.riffreporter.de/de/wissen/chatgpt-ki-reiseplanung-urlaub-tipps
Show Original Post
ycombinator (@ycombinator@rss-mstdn.studiofreesia.com)
The warning signs the AI bubble is about to burst
https://www.telegraph.co.uk/business/2025/08/20/ai-report-triggering-panic-and-fear-on-wall-street/
#ycombinator #ChatGPT #Technology #Meta #Business #Standard #Artificial_Intelligence #Big_tech #Silicon_Valley #US_content #Shares
Show Original Post