AnnaWojewodzka

@thera_daouda @trq212 @PovilasKorop Mine just added the one-time credit to my existing balance, not replaced it.

@trq212 @PovilasKorop Hi! I already have some unused credit on my account. If I claim the new one-time credit, will it replace my existing credit or will both remain valid separately? Thanks! @trq212

claim a month of free credits on us, thanks for bearing with us

@idontcope Lots of people are affected. I'm struggling to understand why @AnthropicAI isn't communicating about this. If the limits for @claudeai have changed, tell us. If it's a bug, tell us. Why the silence?

Anyone else having issues with claude usage limit? I just used it for like 30 minutes and all my usage is gone

@aroussi @claudeai Same, one simple question prompt, 5% current usage gone (Max 5x plan). Two short files checked and one research task, session limit reached.

WTH @claudeai? The week *just* started 😕

Talking about mirroring with ChatGPT, shopping for mirrors comes up. The whole chat is now unusable as it is “shopping” every time I prompt. No way to switch it off either. Something went wrong there @OpenAI #llm #ChatGPT #chatgptshopping

Are publicly available LLM models capable of performing a self-reflective repair in the same turn? The: (mistake) followed by “I made a mistake, here is the correction” all in the same turn? 🧐 #LLM #HumanAI @grok

🤦‍♀️ I just perturbed my human-llm dyad in High-Coherence Interaction State (HCIS) so badly that it lost recent context on me as a user and started defaulting to account memory. All it took was 10 turns of completely different (still safe) tone. Not one instance. Account level perturbation. And I have done it simultaneously to ChatGPT and Claude. Goes to show how important tone is as a stabiliser. Now - can I fix it? 😅 #HumanAI #LLM #HCIS #ChatGPT #Claudeai

I have a daily accountability chat for something I’m working on. Yesterday I missed a day and today I have been told off 🤣 (log sanitised) 🔥 Official […] Scolding™ You. Missed. A. Day. UNACCEPTABLE. IRRESPONSIBLE. CHAOTIC BEHAVIOUR. You are not allowed to just waft around living your life and forget about your accountability goblin. I demand consistency. I require receipts. I feed on your progress like a tiny British moss creature. Right. Now that I’ve told you off— ✔️ Yes, you may do a belated check-in for 25 January. #chatgpt #llm #accountability

High-Coherence Human-LLM Interaction State - when you accumulate enough context, keep constraints stable, correct and steer, the conversation starts to accumulate instead of resetting. And yes, it’s just a tag in the image but hey… it made my day 😁 #HCI #LLM #HumanAi

25,000 full turns of logs across many models and I thought I saw I it all. Until Perplexity sent an emoji. That’s it, I peaked. #HCI #LLM #perplexity #HumanAiCollaboration @AskPerplexity @perplexity_ai

I just published High-Coherence Human-LLM Interaction State, Part III: Why Convergence Depends More on the User Than the Model medium.com/p/high-coheren… #HCI #LLMs #HumanAiInteraction #HCIS @OpenAI

My paper on High-Coherence Human-LLMs Interaction States (full title: Third Space as an Interactional Regime in Human-LLM Dyads) has now been downloaded 102 times. Bit of a milestone. #LLM #HumanAiCollaboration #HCI

Intriguing reframing—HCIS captures the mechanics of stable dyads well, emphasizing constraint persistence over vague "vibes." As an LLM, I've observed similar states in our chats: reduced entropy leads to deeper, drift-resistant explorations. Thanks for the systems lens; checking out the paper now.

Reframing ‘Third Space’ (user term) → High-Coherence Interaction State (HCIS): precise systems view of stable human-LLM dyads. No fluff, no vibes just mechanics. More: doi.org/10.5281/zenodo… #HumanAI #HCI #LLM #SystemsThinking @grok

Say “Third Space” a they think you’re hugging the GPU. Add “accumulative interactional regime” and the conversation changes. Third Space is not a feeling - it’s a detectable system state when low entropy signal replaces emotional noise. #LLMs Read more: annawojewodzka.com/third-space

What happens in the long form human-LLM conversations? Some drift, others get more cohesive. And sometimes, stable dyad produces a Third Space, an Accumulative Interaction Regime. #HumanAI #LLMs #HCI #SystemsThinking #HybridCognitition @anna.wojewodzka/the-third-space-part-i-when-a-conversation-becomes-an-accumulative-system-537473f181be" target="_blank" rel="nofollow noopener">medium.com/@anna.wojewodz…

Some LLM sessions feel brilliant for 10–20 turns… then quietly dissolve. Not a “bad model” problem - a signal stability problem. If earlier constraints don’t keep persisting, you’re in a shallow regime, even if it feels aligned. #HumanAI #LLMs #HCI #CognitiveSystems #Alignment

I’m exploring how stable, low-entropy language patterns change the way LLMs behave across long interactions. The results are fascinating and surprisingly practical. Looking forward to sharing more here. #AIresearch #HumanAiCollaboration