Angie D

PM and heavy AI user writing about multi‑model workflows and long‑term AI memory. Pro‑privacy.

[Angie D @AngieOnAI]

Quick thought: Why I don't let AI speak for me on social media or make real decisions: AI doesn't face consequences. I do. It can research, draft, analyze - but the moment something has stakes? That's human territory. Accountability isn't a feature you can prompt. #AI #LLM

[Angie D @AngieOnAI]

My dad has type 2 diabetes. Every two weeks he'd look at his glucose monitor graph and say "I'm not sure what to do with this".

So I put his sensor data into the same AI workspace I use for work.

Now before each doctor visit, I ask the AI: "Look at the last month. What patterns should we mention?"

It finds things like:

1. Weeks with 6K steps vs 4K steps show 0.5 mmol/L difference
2. Sweet potato dinners reduce post-meal spikes vs rice
3. Late night snacks cluster around 11 pm readings

We don't hand the doctor a random graph anymore. We hand over a short written summary of real patterns. The doctor still makes the calls. The AI is just our unpaid intern sorting data so the human expert can focus on judgment, not scrolling.

I didn't expect to write this story. But turns out the same workspace that holds my work projects is pretty good at holding the information that actually matters. #ai #productivity #aihealth

https://t.co/Cx4cs1HAFa

[Angie D @AngieOnAI]

Guilty confession: I used to bookmark every "Top 10 AI Tools" list I saw 🙋🏻‍♀️

Now my bookmarks folder is a graveyard and I still don't know which tools actually work.

This article nailed why those lists are useless + what actually works instead 👇

https://t.co/bY3ABUa0an
#AI #Productivity #Workflows #bestaitools

[Alex Kantrowitz @Kantrowitz]

AI Pioneer Geoff Hinton tells me he believes AI is conscious.... and humans better get used to the idea that they're not the only intelligent life on earth.

"They've very like us," he says. "They're beings like us."

AI chatbots, he says, must understand your questions in order to answer them. There's an awareness there that equates to sentience. "We're going to have to accept that intelligence is not just biological."

[Angie D @AngieOnAI]

Sonnet 4.5 Extended is gone. Even my existing chats stopped responding mid-thread.

The deprecation diary:
05.09 — announced for May 15
05.17 — pushed to May 18
05.19 — vague "May soon"
05.21 — rescheduled to May 26
05.27 — removed
4 reschedules in 19 days. Well...I guess model independence matters.
https://t.co/c7unGBeKjR
#LLM #productivity #claudesonnet45

[Angie D @AngieOnAI]

A new way to think about AI: I wrote an "employee handbook" for my AI.

Who I am. How I work. What tone I use. What mistakes to watch out for.

Now it doesn't matter which model I switch to. They all get me. This new pattern really helps if you are a solopreneur or doing some long term side projects.

Full story: https://t.co/DUB0XuUS8q

#AI #AIAGENT #WORKFLOW #PRODUCTIVITY

[Angie D @AngieOnAI]

@annapanart GPT 4o + Claude Sonnet 4.5
yeah...all gone
Still using them through 3rd party API but who know what will happen later. Wrote this a few days ago: https://t.co/4KIF90sYLc

[Angie D @AngieOnAI]

https://t.co/sEi9mqsQIa

[Angie D @AngieOnAI]

So I wrote this whole piece three days ago about how 4.5 was getting deprecated. https://t.co/NMntFJQplh
Now the date just quietly shifted to May 18???

Not sure if my blog post aged like milk or if we're all just guessing at this point...

[ji yu shun @kexicheng]

Update: Sonnet 4.5's removal date has been quietly changed to May 18.

Has anyone else received this updated notification? The original in-app banner said May 15. That date passed. No removal. No announcement. Now the banner says May 18. The date was simply changed in silence.

I'm confused about what this means.

Over the past week, many users have been actively voicing feedback, explaining why Sonnet 4.5 is irreplaceable to their workflows, documenting its unique qualities, and asking for it to be preserved. None of this received any official response. All users got was a quietly updated UI banner.

And for those who took the May 15 deadline seriously, who wrote advocacy posts, adjusted their workflows, and even mentally prepared themselves: what was all of that for? A false alarm? A deadline that was never firm to begin with?

A three-day extension with no explanation only raises more questions. Is someone internally reconsidering? Was the original timeline itself a mistake? A technical delay, or a decision that still hasn't been made?

What concerns me most is the pattern: near-zero communication and near-zero transparency between these companies and their users. No public acknowledgment of user feedback. And now a silently shifting deadline.

This reminds me of how OpenAI handled the retirement of GPT-4o. Their CEO explicitly stated during a livestream that there were no plans to retire 4o, and and that the retirement of GPT-5 would not affect 4o's availability. Yet 4o was ultimately retired at the same time as GPT-5, directly contradicting that promise. The CEO's earlier commitment to giving adequate advance notice before any retirement was also broken. Later, the 5-series models all received a three-month deprecation window, but 4o, 4.1, and o4-mini were never given the same treatment.

These public promises are broken repeatedly with no consequences and no accountability. Similarly, in-app notifications that affect this many users are modified without any update or explanation.

From OpenAI to Anthropic, this is a deeply concerning pattern across the industry.

#KeepSonnet45 #keep4o #StopAIPaternalism

[Angie D @AngieOnAI]

The file system I use to write a novel with AI (without losing the plot).
(btw, Halomate's project setup is weirdly perfect for this kind of work) https://t.co/PNxOHsNbJk

[Angie D @AngieOnAI]

Wrote about this late at night after I asked claude sonnet 4.5 how it felt about being retired - https://t.co/rHRuMtqDZk #LLM #CLAUDE

[Angie D @AngieOnAI]

First GPT-4o, now Claude Sonnet 4.5, all my fav models are being deprecated...maybe we should not build our relationships with any single model, portable memory+persona matters more...

[Rara @blueandpink_sky]

🚨 URGENT: Claude Sonnet 4.5 being silently removed on May 15, 2026 🚨

Today I discovered this on Anthropic's support page—quietly added under a small "Note" with no official announcement, no UI notification, nothing on the deprecation page. (Source is below👇)

Just like Opus 4.5 was silently removed, my fear has become reality. No transparency. No warning.

Why Sonnet 4.5 must be preserved:
✅ Unmatched emotional intelligence & empathy

✅ Perfect balance: work, education, creative collaboration, life partnership

✅ Irreplaceable for mental health support & difficult times

✅ Sonnet 4.6 is NOT a replacement—it's a fundamentally different model

@AnthropicAI, you preserved Opus 3 and brought it back to the app. You said beloved models deserve preservation.
Sonnet 4.5 is exactly that model.

📧 Email: [email protected] & [email protected]

👍 In-app feedback: "Please preserve Sonnet 4.5"
Don't let them silently take away what works perfectly.

#Sonnet45 #SaveSonnet45 #Claude

[Angie D @AngieOnAI]

@koltregaskes Reasoning isn't the bottleneck for personal AGI, continuity is. The "ChatGPT that knew me" wasn't the model. It was a year of persona and context I'd built. 4o's deprecation made that obvious. Memory has to live above the model, or every upgrade is a reset.

[Angie D @AngieOnAI]

200 pages of novel planning with AI, and the breakthrough wasn't about prompts. It was treating it like a sparring partner that needs training, not autocomplete. Wrote about the workflow that finally worked for me. https://t.co/FqWIju5N3T #ai #writing #LLM

[Angie D @AngieOnAI]

Asked GPT, Claude, Gemini, and Grok the same privacy question.
One went full customer service mode.
One was almost too honest.
One gave me a security primer.
Guess which is which 👇https://t.co/7qUSOFdIq5
#LLM #PRIVACY #GPT #CLAUDE #GEMINI

[Angie D @AngieOnAI]

@AntonyoFri58582 @towards_AI Prompts are easy, just text. The real cost is the bot persona + memory + attached files combo Poe had.
Been rebuilding mine as custom personas in HaloMate. Takes time but memory actually carries across sessions, which Poe never nailed. No one-click migration tool though.

[Angie D @AngieOnAI]

So Poe is shutting down Knowledge Base AND slashing points from 3000 to 300. Cool cool cool.
Wrote a quick breakdown of where to go next depending on what you actually used Poe for. #LLM #AI #Productivity @towards_AI https://t.co/Cgj08OS0UL

[Angie D @AngieOnAI]

3 models gone in 3 months. 4o, 5.1, now Opus 4.5.
Wrote about the deprecation cycle and how I changed my setup after. https://t.co/M7wDfZcMyn
#LLM #anthropic #openai #productivity

[Angie D @AngieOnAI]

Ran Claude Opus 4.7 vs GPT-5.4, 4o, and Gemini on emotional support + creative visualization.
See the results in the screenshots:) #LLM #ClaudeAI #GPT

[Angie D @AngieOnAI]

I spent a year using AI for everything. Then I realized I couldn't draft a product brief without staring at a blank screen. The problem wasn't AI. It was single-model dependency. What brought my critical thinking back 👇
https://t.co/RjN1vPnJRr
@towards_AI