Anthropic's Claude agents show improved performance with 'dreaming' feature

COVERAGE [37]

1. Simon Willison TIER_1 Dansk(DA) · 2026-05-06 15:58

   Live blog: Code w/ Claude 2026

   <p>I'm at Anthropic's Code w/ Claude event in 2026, and I'll be live blogging the keynote and a few other notes throughout the day.</p> <p>Tags: <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a>, <a href="…

2. Mastodon — sigmoid.social TIER_1 · [email protected] · 2026-05-11 17:41

   Everything is so weird. # ai # llm

   Everything is so weird. # ai # llm

3. Medium — Claude tag TIER_1 · Chew Loong Nian - AI ENGINEER · 2026-05-09 08:29

   I Let Claude Dream for 4 Hours — Today’s Agent Just Killed Yesterday’s by 5.4× on 18 Repeat Tasks

   <div class="medium-feed-item"><p class="medium-feed-image"><a href="https://pub.towardsai.net/i-let-claude-dream-for-4-hours-todays-agent-just-killed-yesterday-s-by-5-4-on-18-repeat-tasks-0d09741fe6f1?source=rss------claude-5"><img src="https://cdn-images-1.medium.com/max/1108/1*…

4. Mastodon — sigmoid.social TIER_1 · [email protected] · 2026-05-06 11:26

   # TokenGenerators are not the future. They are an evolutionary dead-end. # LLM is not # AI

   # TokenGenerators are not the future. They are an evolutionary dead-end. # LLM is not # AI

5. Mastodon — sigmoid.social TIER_1 · [email protected] · 2026-05-05 01:54

   A) Citing is fundamental & necessary to STEM - no matter the source. That verifiable citations aren't supported is anti-education B) Student/instructor relation

   A) Citing is fundamental & necessary to STEM - no matter the source. That verifiable citations aren't supported is anti-education B) Student/instructor relationships have always been based on- & enforced by- professionality & trust. To ignore that relationship is profoundly disre…

6. Mastodon — sigmoid.social TIER_1 · [email protected] · 2026-05-05 01:53

   Canadian universities have agreed to allow a tech company to provide an "in-house" AI / LLM service to all its students and staff. As an academic, I'm disappoin

   Canadian universities have agreed to allow a tech company to provide an "in-house" AI / LLM service to all its students and staff. As an academic, I'm disappointed students won't be able to provide verifiable citations to their AI "prompting-and-statistical-response-generation" p…

7. Mastodon — sigmoid.social TIER_1 · [email protected] · 2026-05-02 18:52

   What’s Claude’s last name? # Claude # AI # Anthropic

   What’s Claude’s last name? # Claude # AI # Anthropic

8. Mastodon — fosstodon.org TIER_1 · [email protected] · 2026-05-09 07:56

   RE: https:// loops.video/v/dTLtGXCtBL This is why the current round of LLM models are not in any way AI. They are not intelligent, nor are they sentient. Please

   RE: https:// loops.video/v/dTLtGXCtBL This is why the current round of LLM models are not in any way AI. They are not intelligent, nor are they sentient. Please reflect on your AI Psychosis and seek help if you are in difficulties 🙂🖖 # AI # LLM # ChatGPT

   LINKS loops.video/…/dTLtGXCtBL
9. Mastodon — fosstodon.org TIER_1 · [email protected] · 2026-05-06 21:41

   RE: https:// mathstodon.xyz/@internic/11651 8341455538394 One thing that seems to confuse discussions of the capabilities of LLMs is that we're heavily biased t

   RE: https:// mathstodon.xyz/@internic/11651 8341455538394 One thing that seems to confuse discussions of the capabilities of LLMs is that we're heavily biased to think of them as human-like (as I expanded on in the quoted post). So here's a question for which Fedi seems well suit…

   LINKS arxiv.org/…/2603.09678
10. Mastodon — fosstodon.org TIER_1 · davep · 2026-05-04 20:04

    Thinking about how I wouldn't have started from here, to get to where I want to go: https:// blog.davep.org/2026/05/04/i-wo uldnt-start-from-here.html # ai # ll

    Thinking about how I wouldn't have started from here, to get to where I want to go: https:// blog.davep.org/2026/05/04/i-wo uldnt-start-from-here.html # ai # llm # agent # python # programming

    LINKS blog.davep.org/…/i-wouldnt-start-from-her…
11. Mastodon — fosstodon.org TIER_1 · [email protected] · 2026-05-04 03:08

    Guys what’s your opinion on this. Should junior software engineers learn to be great at writing software Or Be great at writing software with llm assistance? On

    Guys what’s your opinion on this. Should junior software engineers learn to be great at writing software Or Be great at writing software with llm assistance? One is reading a lot of code and the other is writing code # programming # software # ai # claude

12. Mastodon — fosstodon.org TIER_1 · [email protected] · 2026-05-04 01:42

    This bullshit is almost worse than LLM slop. I never liked the guy, even though I’m an atheist/agnostic/non-religious, but this is a level of irony that I can’t

    This bullshit is almost worse than LLM slop. I never liked the guy, even though I’m an atheist/agnostic/non-religious, but this is a level of irony that I can’t comprehend. He’s treating an algorithm, the LLM, as if it were a kind of God. Is AI the next phase of evolution? - UnHe…

13. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-11 10:39

    @ fdroidorg exciting updates. is there any kind of labelling or categorisation in # fdroid of whether an app is made with # AI / # LLM code generator or not?

    @ fdroidorg exciting updates. is there any kind of labelling or categorisation in # fdroid of whether an app is made with # AI / # LLM code generator or not?

    LINKS floss.social/@fdroidorg
14. Mastodon — mastodon.social TIER_1 · ngate · 2026-05-09 00:25

    Can # LLMs model real-world systems in TLA+? 🤔 Oh sure, because who wouldn't want a linguistically confused # AI applying # logic it doesn't truly comprehend to

    Can # LLMs model real-world systems in TLA+? 🤔 Oh sure, because who wouldn't want a linguistically confused # AI applying # logic it doesn't truly comprehend to complex systems modeling? 🙄 Next up: teaching your cat to do your taxes! 🐱💼 https://www. sigops.org/2026/can-llms-model…

15. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-08 22:32

    🤬 Am I the only one who wishes F-Droid had an anti-feature toggle for "AI" (LLM slop)? # FDroid # AI # slop # LLM # ArtificialIntelligence # Android

    🤬 Am I the only one who wishes F-Droid had an anti-feature toggle for "AI" (LLM slop)? # FDroid # AI # slop # LLM # ArtificialIntelligence # Android

16. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-06 01:25

    AI (LLM) only gives value if you accept the irrational framing of capitalist economics and society. #LLM #AI

    AI (LLM) only gives value if you accept the irrational framing of capitalist economics and society. #LLM #AI

17. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-05 19:25

    (11/12) How does the bubble unwind? Already is. Watch the pricing games and the shift to charging for tokens. AI companies have been operating on Loony Toons ph

    (11/12) How does the bubble unwind? Already is. Watch the pricing games and the shift to charging for tokens. AI companies have been operating on Loony Toons physics, but gravity is catching up. As the bubble pops, expect the house of cards around data center financing to implode…

    LINKS misaligned.markets/what-to-expect-ai-bubb… misaligned.markets/what-to-exp
18. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-05 19:25

    (7/12) Just as LLMs rely on tooling and governance, market systems rely on rules and laws. There is no “emergent” LLM intelligence, just as there is no “self-re

    (7/12) Just as LLMs rely on tooling and governance, market systems rely on rules and laws. There is no “emergent” LLM intelligence, just as there is no “self-regulating” free market. Both are myths of autonomy. https:// misaligned.markets/ai-hype-mar ket-fundamentalism/ # LLM # n…

    LINKS misaligned.markets/ai-hype-market-fundame… misaligned.markets/ai-hype-mar
19. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-05 19:25

    (5/12) Specifically, LLMs benefit from dark patterns that “launder” the idea that a mind is on the other end of the screen. This obfuscates the intentions of AI

    (5/12) Specifically, LLMs benefit from dark patterns that “launder” the idea that a mind is on the other end of the screen. This obfuscates the intentions of AI companies while making invisible the labor of the people lower in the chain. I don't think there's a “neutral” version …

    LINKS misaligned.markets/llm-intelligence-is-a-… misaligned.markets/llm-intelli
20. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-05 19:25

    (2/12) To understand the bubble, we first have to demystify the tech. LLMs require staggering amounts of human involvement during training and at runtime. Any “

    (2/12) To understand the bubble, we first have to demystify the tech. LLMs require staggering amounts of human involvement during training and at runtime. Any “intelligence” is heavily scaffolded. That most discussions don't start with this notion criminal. https:// misaligned.ma…

    LINKS misaligned.markets/understanding-large-la… misaligned.markets/understandi
21. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-05 19:25

    (1/12) The 'AI bubble' is a political economy story I've spent over 20k words unraveling. LLMs aren't just a technology; they're a mirror of the same delusions

    (1/12) The 'AI bubble' is a political economy story I've spent over 20k words unraveling. LLMs aren't just a technology; they're a mirror of the same delusions that drive market fundamentalism: a naive understanding of innovation, deference to capital, and the denigration of huma…

22. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-03 21:51

    Very accessible and valuable critique of the # AI subspecies known as # LLM . Will human minds still be special in an age of AI? https://www. theguardian.com/bo

    Very accessible and valuable critique of the # AI subspecies known as # LLM . Will human minds still be special in an age of AI? https://www. theguardian.com/books/2026/may /03/will-human-minds-still-be-special-in-an-age-of-ai?CMP=Share_AndroidApp_Other

23. Mastodon — mastodon.social TIER_1 Nederlands(NL) · [email protected] · 2026-05-03 13:30

    @joepbc whoever leans too much on #llm #ai quickly becomes dumber.

    @ joepbc wie teveel op # llm # ai leunt wordt al snel dommer.

24. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-02 17:21

    Asked claude for options on marking parts of songs on https:// null.dangerthirdrail.com It output an HTML file with these options: This is unreal! Not "Claude D

    Asked claude for options on marking parts of songs on https:// null.dangerthirdrail.com It output an HTML file with these options: This is unreal! Not "Claude Design" or whatever, regular Claude Code. The takeaway IMO is that well specified instructions can take you very far. # A…

    LINKS null.dangerthirdrail.com
25. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-02 15:35

    "Any sufficiently advanced technology is indistinguishable from magic" is hitting pretty hard in the age of # LLM and # AI

    "Any sufficiently advanced technology is indistinguishable from magic" is hitting pretty hard in the age of # LLM and # AI

26. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-02 13:30

    This is what we're using # AI for now, is it? https:// ldstephens.net/posts/claude-sa ved-me-hours-of-link-cleanup/ "Three minutes with Claude saved what would

    This is what we're using # AI for now, is it? https:// ldstephens.net/posts/claude-sa ved-me-hours-of-link-cleanup/ "Three minutes with Claude saved what would have been a couple of hours of manual work." Or, about 15 seconds in an editor (any editor) with search/replace. Or abou…

27. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-05-02 10:54

    🎥 When AI Gets History Wrong 🔗 https://www. youtube.com/watch?v=NaeQm2KYjac For notes associated with the # video in # pdf format: https:// jamestabor.com/wp-co

    🎥 When AI Gets History Wrong 🔗 https://www. youtube.com/watch?v=NaeQm2KYjac For notes associated with the # video in # pdf format: https:// jamestabor.com/wp-content/uplo ads/2026/04/Chat-on-Pantera.pdf # History # Histodons # AI # ArtificialIntelligence # Technology # Tech # Res…

    LINKS youtube.com/watch
28. Mastodon — mastodon.social TIER_1 · aihaberleri · 2026-04-30 04:24

    📰 LLM Safety Crisis: 54.4% of AI Health Robots Violate Medical Ethics (2026) A new study benchmarks the safety of large language models for robotic health atten

    📰 LLM Safety Crisis: 54.4% of AI Health Robots Violate Medical Ethics (2026) A new study benchmarks the safety of large language models for robotic health attendants, revealing alarming violation rates. Over half of models failed ethical safeguards, raising urgent concerns for cl…

29. Mastodon — mastodon.social TIER_1 Türkçe(TR) · aihaberleri · 2026-04-30 04:24

    📰 Safety of Large Language Models in Robotic Medical Assistants: Critical Risks in 2026 Large language models determining the safety of robotic healthcare assistants

    📰 Büyük Dil Modellerinin Robotik Hekim Yardımcılarında Güvenliği: 2026'da Yaşanan Kritik Riskler Robotik sağlık yardımcılarının güvenliğini belirleyen büyük dil modellerinin etik ve teknik sınırları yeni bir araştırma dalgasıyla ortaya çıkıyor. Üç farklı akademik kaynaktan derlen…

30. Mastodon — mastodon.social TIER_1 · aihaberleri · 2026-04-30 04:24

    📰 2026 Human-in-the-Loop LLM Benchmark: Why Smaller Models Outperform Giants in Math Assessment A new study benchmarks large language models in automating secon

    📰 2026 Human-in-the-Loop LLM Benchmark: Why Smaller Models Outperform Giants in Math Assessment A new study benchmarks large language models in automating secondary math competency assessment, revealing that architectural design outweighs model size in rubric-based tasks. Human-i…

31. Mastodon — mastodon.social TIER_1 Türkçe(TR) · aihaberleri · 2026-04-30 04:24

    📰 Measuring Math Competencies with Human-in-the-Loop Benchmarking (2026): AI and Teacher Collaboration... Middle School Math Competencies of Artificial Intelligence Models

    📰 Human-in-the-Loop Benchmarking ile Matematik Yetkinliklerini Ölçmek (2026): AI ve Öğretmen İş Bir... Yapay zeka modellerinin ortaokul matematik yetkinliklerini değerlendirmesinde insan katılımının kritik rolü ortaya çıkıyor. Yeni bir araştırma, bu sistemin nasıl daha adil ve do…

32. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-04-28 15:40

    If you're attaching probabilistic machines to your most important business functions, and then being surprised when they break your business, perhaps you need t

    If you're attaching probabilistic machines to your most important business functions, and then being surprised when they break your business, perhaps you need to be removed from management, because you have failed a *very* simple test of clear thinking. # AI https://www. tomshard…

    LINKS tomshardware.com/tech-industry
33. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-04-28 12:51

    Sometimes it's convenient when an agent is missing something and still manages to get it done by reinventing the universe. https://www. tomshardware.com/tech-in

    Sometimes it's convenient when an agent is missing something and still manages to get it done by reinventing the universe. https://www. tomshardware.com/tech-industry /artificial-intelligence/claude-powered-ai-coding-agent-deletes-entire-company-database-in-9-seconds-backups-zapp…

    LINKS tomshardware.com/tech-industry
34. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-04-28 04:41

    I used to think it's just recklessness to give an # AI complete control over your projects and risk deleting something important At this point, I think these pe

    I used to think it's just recklessness to give an # AI complete control over your projects and risk deleting something important At this point, I think these people are just morons https://www. tomshardware.com/tech-industry /artificial-intelligence/claude-powered-ai-coding-agent…

    LINKS tomshardware.com/tech-industry
35. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-04-27 22:19

    The machines are not, in fact, fine. - Boxo Barks https:// boxobarks.leaflet.pub/3misaejn oqs2k > "Claude, write me a viral debut blog post using these three so

    The machines are not, in fact, fine. - Boxo Barks https:// boxobarks.leaflet.pub/3misaejn oqs2k > "Claude, write me a viral debut blog post using these three sources, make no mistakes." # AI # Claude

    LINKS boxobarks.leaflet.pub/3misaejn
36. Mastodon — mastodon.social TIER_1 · [email protected] · 2026-04-27 21:30

    I'm very much torn between “use an LLM ‘agent’ and get what you deserve,” and blaming the tech industry and the overly-credulous press for hyping these chatbots

    I'm very much torn between “use an LLM ‘agent’ and get what you deserve,” and blaming the tech industry and the overly-credulous press for hyping these chatbots so much https://www. tomshardware.com/tech-industry /artificial-intelligence/claude-powered-ai-coding-agent-deletes-ent…

    LINKS tomshardware.com/tech-industry
37. Mastodon — mastodon.social TIER_1 · ScottE · 2026-04-27 18:56

    On the flight from Munich to Malta, the guy in the seat in front of us was watching YouTube videos at 2x speed and taking furious notes about "how to use Claude

    On the flight from Munich to Malta, the guy in the seat in front of us was watching YouTube videos at 2x speed and taking furious notes about "how to use Claude to build a business and make millions". (Seriously, that was one of the videos). My dude, if your plans are basically "…