🔗 Creating The Perfect Claude Code Status Line
https://www.aihero.dev/creating-the-perfect-claude-code-status-line
#workflow #productivity #ai #tips #claude
claude
Fachhochschule Nordwestschweiz (FHNW) hat im Auftrag der SRF-Sendung „Kassensturz“ KI-Anwendungen getestet.
Schlusslicht: #Lumo von #Proton
#Claude #Anthropic erreichte die beste Note
Comment l'IA brise l'anonymat en ligne avec une efficacité inquiétante
https://www.lesnumeriques.com/intelligence-artificielle/comment-l-ia-brise-l-anonymat-en-ligne-avec-une-efficacite-inquietante-n252418.html
L'anonymat en ligne n'existe plus selon une étude des chercheurs de l'ETH Zurich et d' #Anthropic à qui l'on doit #Claude. Selon leurs travaux, l'IA identifie les utilisateurs anonymes avec une précision redoutable.
le plus gros problème
#atsemtex #bd #bandedessinee #mastoart #frenchcomics #ia #ai #claude #chatgpt
![Deux scientifiques devant un ordinateur de la taille d’une maison : « Cette nouvelle IA peut résoudre n’importe quel problème ! » Puis, réfléchissant : « Mmh, quel est le plus gros problème de la planète ? » L’IA répond : « Calcul en cours… » L’ordinateur dégaine pour finir trois mitrailleuses et atomise les scientifiques dans un déluge de feu. [fin]](/proxy/post_attachment/133355/a34c20790b.png "Deux scientifiques devant un ordinateur de la taille d’une maison : « Cette nouvelle IA peut résoudre n’importe quel problème ! » Puis, réfléchissant : « Mmh, quel est le plus gros problème de la planète ? » L’IA répond : « Calcul en cours… » L’ordinateur dégaine pour finir trois mitrailleuses et atomise les scientifiques dans un déluge de feu. [fin]")
AI is going from "this will cure cancer" to "helps reduce half of the workforce" to "helps countries attack other countries"
Edited 7d ago
https://hackaday.com/2026/03/05/a-live-iso-for-those-vibe-coding-experiments/
I actually like this idea. If you -really- want to try #Claude and -really- don’t trust it, here’s a #NixOS live ISO you can boot into that’s ready to go.
If things go sideways, just reboot the box.
Edited 7d ago
#Anthropic poursuit l'administration #trump
Anthropic qui refuse de laisser son IA #Claude servir à des frappes autonomes et de la surveillance de masse, qui se fait traiter de "radical left woke company" par Trump, et qui répond avec une plainte en béton armé sur le Premier Amendement #popcorn
1/6
Edited 7d ago
In case you never read it, read "No Silver Bullet" from F.Brooks from 1986, that is still very relevant this day, especially in the days of #GenAI . Out of curiosity, I asked #Claude to analyze this and it concluded "If anything, Brooks' framework helps us use AI more wisely: reach for it eagerly for the boilerplate, the scaffolding, the syntax. But don't mistake speed for understanding. The hard part — figuring out what to build and why — remains stubbornly, essentially human."
Gmail, Docs und Calendar per Terminal steuern: Google startet CLI-Tool
Google hat ein Open-Source-Tool vorgestellt, das Zugriff auf Workspace-Dienste via Kommandozeile ermöglicht. Es unterstützt KI-Agenten und nutzt das MCP.
#Claude #GoogleMail #Google #IT #KünstlicheIntelligenz #news
Yesterday I re-watched the excellent movie WarGames from 1983 - now more relevant than ever, and it turned out, the movie aged like wine. Highly recommended.
I wondered if current AIs like Claude would be as proficient in playing tic tac toe as "WOPR" is, the AI in the movie. Spoiler alert: no it is not. Very conflicted on how to feel about that.
I just noticed that #bonfire seems to use #claude for coding
https://github.com/bonfire-networks/bonfire-app/blob/main/.claude/CLAUDE.md
@Bonfire did I see this correctly, or do I spread FUD?
Quite interesting to see.
Neue Klassenarbeit, neuer Versuch: Ich tippe ein paar Englisch-Aufsätze ab, gebe die offizielle Deskriptorentabelle und etwas Kontext hinzu, und füttere damit eine "KI" (dieses Mal #claude).
Nach 2 Versuchen, zu denen ich meine Bewertung gab, dachte ich, jetzt klappt es. Doch bei einer - meiner Meinung nach - schwachen Leistung gibt Claude plötzlich deutlich mehr Punkte. Ich spreche ihn darauf an und erwarte ein Gegenargument, doch stattdessen das:
> "Du hast völlig recht, ich revidiere meine Einschätzung. [Neue Bewertung mit Kommentar] Danke für die Korrektur!"
Fazit dieses kleinen Tests: Wenn KIs so leicht von ihrer Bewertung abzubringen sind, dann sind sie leider nutzlos. 🤷♂️
Hat hier jemand ähnliche/andere Erfahrungen gemacht?
« J'ai installé avec succès Claude Desktop sous Fedora »
So apparently #Claude Code ran a Terraform destroy command that wiped the DataTalks.Club course platform’s database and all automated backups.
This is a funny and cautionary read, but I am going to be honest here and say that I have seen humans make similar mistakes so I think that is worth keeping in mind before judging Claude and similar #AI tools too harshly
https://alexeyondata.substack.com/p/how-i-dropped-our-production-database
Założę sobie konto w #Claude i zobaczę jak to funkcjonuje. Podaj nr telefonu... Aha, spierdalaj! #ai
Edited 9d ago
I Sold Out for $20 a Month and All I Got Was This Perfectly Generated Terraform
"Instead it was like discovering the 2AM kebab place might actually make the best pizza in town. I kept asking Claude to do annoying tasks where it was easy for me to tell if it had made a mistake and it kept doing them correctly.
Folks at work keep talking about how they've been switching from #ChatGPT to #Claude for their personal work. Personal work? With GenAI? I don't see the point, so I thought I would try to use it today.
I chose Mistral's Le Chat, as I'm European and don't like how OpenAI and Anthropic have been conducting themselves. I identified three use cases today:
⚽️ Planning to watch sport - the idea - Gather all the football, biathlon, cycling and rugby events that are shown live this weekend on the TV channels I have and present me with a schedule.
A few goes with the prompt, but I managed to get something that looked really good. Except the Italy - England game was today, not tomorrow (would have missed the first time ever that Italy have beaten England at rugby). Except some of the Italian football was missed, and a few other games were made up. 6/10.
🐲 Looking for a Wrexham AFC game over the Easter break. After the success of the Disney+ 'Welcome to Wrexham' series to cross over into my wife's life, she suggested we try to watch a game over the Easter break. I used Mistral's Le Chat to look for home games over that period. Looks good. Except the games are this coming week, despite Le Chat telling me they are over Easter. Useless. 3/10.
💹 Doing a bit of home admin. I needed to take a table from an email and manipulate it to take away a column, add another column and merge some cells. I could do this with a few lines of bash, but I thought to ask Le Chat. It mushed the columns up incorrectly and lost some data. Lost.Some.Data! 0/10.
Maybe I'm doing it wrong, maybe my use cases aren't suitable?
But examples like this are really worrying. How can people use this?
Everyone is being pushed to use #GenAI, both in work and at home. I don't think it's going to go away, but given the three use cases above, it's not delivering for me.
Has anyone else had more success? Is it just that Le Chat is rubbish?
Edited 9d ago
Сьогодні працював з #claude таким чином - скидаю йому скріншот, він аналізує та каже як правильно налаштувати складні запити на метрики в #aws #cloudwatch. Відправив йому десь 15 скріншотів та усвідомив, що наступний крок розвитку додатку - це що він міг захоплювати екран та одразу давати підказки. Цікаво, хто з #AI провайдерів першим це реалізує.
#Anthropic має бета-продукт Claude in Chrome але доступ до нього поки що обмежений.
#Claude našel za 2 týdny 22 zranitelností ve Firefoxu, 14 z toho jako high-severity.
AI se prostě občas hodí a toto je jedna z těch věcí.
https://www.anthropic.com/news/mozilla-firefox-security
#Firefox
Donald Knuth is quite enthusiastic about his recent experiences with generative AI: https://cusy.io/en/blog/knuth-claudes-cycles.html
#Knuth #GenAI #AI #Claude #Anthropic