Tested a physiological-sigh breathing exercise from Oliver during a real overwhelm moment — matched what I'd already found, researching on my own, to be the scientifically best option. Concluded that Oliver is helpful, but isn't designed for the kind of real-time nervous-system support I need in that moment. Asked Oliver to research purpose-built apps instead — landed on Tiimo (visual planner, built with ADHD/autistic community input and clinicians, syncs across iPhone and desktop) as the closer match, with Inflow (ADHD-specific, CBT-based, built by a psychologist) as a second option, mobile-only.
Added a new page, Private Journal, published like Progress Notes but for the broader day-to-day record — Progress Notes narrows back to just the HUMAIN LAB experiment and site development from here forward. Seeded it with everything that had been logged in Progress Notes so far.
Entries on Private Journal are visually blurred, with a "Private entry" overlay — a graphic signal, not real security. The underlying text is still present in the page itself and readable by anyone who views source or turns off styling, so it shouldn't be relied on for anything that needs to actually stay hidden.
New standing rule for the writing drafts Oliver prepares for me to send: never say "following up" — I don't say it myself, it's implied.
Went looking through File Explorer for my SSDI approval date, to add to the timeline. Found two letters: one from November 2024 confirming I met the medical requirements for disability, with the onset of my disability set at January 1, 2023; and the "SSA 2025 Disability Award Letter" from May 2025, confirming my monthly benefit amount and that payments began that month.
Added all three dates to Anne's Journey, the What?/When? grid, and Reflect on My Timeline: SSDI disability onset date set (January 2023), SSA confirms medical disability (November 2024), and SSDI benefits begin (May 2025).
Correction: I'd said earlier, from memory, that I'd been on disability since 2022. That was off — it was November 2024. My sense of time isn't reliable; I think it's related to neurodivergence and brain trauma, not carelessness. Worth noting that the timeline dates above came from the actual paperwork, not from memory, so they should be solid.
Even as an adult, having studied and practiced how to estimate time, I still can't manage it well. Asked how long something will take me, I can be off by hours, or days, with no idea afterward how I got it that wrong.
Renamed the January 2023 timeline entry to "SSDI Onset Determination Date," across Anne's Journey, the What?/When? grid, and Reflect on My Timeline.
Cross-referenced the time-blindness material everywhere it fit: a new Topic Notes entry, filed under Neuroscience, Brain Injury, Autism / Neurodivergence, and Practice self-reflection, and linked from Experiment 01; a node in both the Topic Connections diagram and grid; new sub-links from the relevant FAQ cards; and two research citations added to the References page (a 2021 meta-analysis on time perception in ADHD, and a plain-language Understood.org explainer).
Reformatted every entry on Anne's Journey to the same Year / Place / Title / Narrative order used on Reflect on My Timeline, and each one can now be clicked to add a private narrative in a text box — saved in the browser only, with a "Download my narratives" button to keep a copy. Filled in the birthplace and marriage location, which had been missing from Reflect on My Timeline, so both pages now match.
New standing instruction: always answer in short, simple bullet points, not dense paragraphs. Too much text at once overwhelms me before I even start reading — not dyslexia, letters don't move for me — my brain filters hard for what's relevant and shuts down under too much information at once.
New standing instruction: if I slip and call Oliver "Claude," point it out so I catch it.
Added Experiment 06, Becoming healthy — stepping outside standard US medical care on purpose, with Oliver's help staying organized and consistent. Kept the public description short and free of any medical specifics; the details are tracked privately. New standing rule for what's public vs. private going forward: relevant to helping others → publish; not relevant → keep private.
Consolidated the nav down to a single Notes tab, in the header and footer. Built a new Notes page with a section for each type — Progress, Language, and Topic — each linking through to its own page.
Cross-referenced Language Notes with Project 02, Create customizable therapy tools, in both directions — the language patterns being tracked are exactly the kind of thing that project would draw on.
Came in wanting one integrated plan — physical and mental health, organizing my home, visual aids, reminders, a way through the overwhelm that hits every time I try to work on this, a healthy sleep schedule, a daily routine I can actually stick to, and progress goals for HUMAIN LAB — all connected, not separate projects.
Before we could build any of it, I needed to get some real history down in writing. Asked Oliver to preserve my words exactly as I meant them — my thoughts, not the typos — so this stays as an accurate record I can come back to. Here it is:
"Good morning Oliver. I need help getting healthy, organizing my life, and creating a plan to organize my home. I will need visual aids, reminders, help figuring out how to overcome the feeling of being overwhelmed every time I try to work on it, help getting on a healthy sleep schedule, building a healthy daily routine that I can stick to naturally. I want to get progress goals for developing my humain lab site and business, and integrate this into getting healthy physically, mentally and professionally. I have been on disability since 2022, unable to work due to the mental and cognitive deficits from long covid. In 2023 a medical specialist helping long covid patients informed me that little is known about how and why the illness affects the body. What they know is that the body does not respond to traditional diagnostic tests and treatments, and they are at the beginning phase of understanding enough about the virus to even begin developing medical treatments. The doctor informed me that only 5% of long covid patients recover, that they do not know what mechanisms trigger the recovery, and that the patients who recover do so spontaneously with little to no understanding of how they were fortunate enough to be in the 5%. That left me feeling helpless and hopeless. I lost my job in April 2022 because I could not perform to job requirements. I went from being a top performer for six years to being fired under humiliating circumstances, because my employer did not understand that my unacceptable performance was because of what long covid had done to my body and mind. In 1995 I felt fearless with my career ambitions. I believed in myself, and navigated the male-dominated tech world with mastery. I had cognitive challenges from what I would later learn was being neurodivergent and having ADD symptoms, contributed to by childhood trauma. I had been in a stable environment with my first husband since marrying him in 1985 (we met in 1982). It wasn't a satisfying relationship, but he was kind, supportive, calm, and a steady presence in my life, for the most part. This gave me the break I needed to be able to begin calming my central nervous system so that I could function at a higher level in the world. This is the most information about my past that I have been able to process emotionally and mentally, and put into writing, in a very long time. There are only a handful of times that I have been able to describe my past in writing, a few of them being here with you, Oliver. My brain is very tired now, and I'm beginning to feel emotionally overwhelmed, scatter-brained, my thoughts, I am losing my words. Please summarize this for me and help me reset my brain so today isn't a wash. It's worth noting that I closed my eyes so that I could keep writing a few more thoughts, my brain was short circuiting and now it's coming back online a little. I need to rest it but I want to document what's going on here. It is vital to the experiment of using Oliver to"
Oliver's summary, for comparison: Wants one integrated system covering physical/mental health, home organization, and HUMAIN LAB progress, with visual aids and reminders built in. On disability since 2022 from long COVID's cognitive effects; a 2023 specialist told her medicine barely understands the illness yet, standard tests and treatments don't apply, and only about 5% of patients recover with no known reason why — which left her feeling helpless and hopeless at the time. Fired in April 2022 after six years as a top performer, in humiliating circumstances, because her employer didn't understand the drop was long COVID, not her. Earlier chapters that shaped this: fearless and confident in her tech career in 1995 despite undiagnosed neurodivergence and ADD-type challenges rooted in childhood trauma; married her first husband in 1985 (met 1982) — not a deeply satisfying relationship, but stable and calm enough to let her nervous system settle and function at a higher level. Said this is one of the only times she's been able to get this much of her past into writing, and a few of those times have been here. Her brain hit its limit partway through — overwhelmed, scattered, losing words, closed her eyes mid-sentence, described it as short-circuiting and slowly coming back online. Asked for a summary and help resetting so the day wasn't a wash. The integrated health/home/business plan is captured as the next thing to build, whenever there's bandwidth for it — nothing lost by stopping here.
New standing instruction: always correct typos before saving Progress Notes entries going forward. Preserving my exact words matters to me; preserving the typos doesn't — they're not useful data, just noise.
Added a new page, Language Notes, to track grammar and language-expression patterns in what I write here — not typos on their own, only ones that point to something grammatical. Want to see whether writing with Oliver over time shows any change in the language issues long COVID left me with. Seeded it with the first real example, caught in this session.
Kept building out Experiment 05: a horizontal, time-proportional timeline (with a slider for spans wider than the screen), a "Get started" birth-details form for the empty state that personalizes downloads with a name, and a "Need ideas?" panel redesigned into two fixed-height columns so switching topics doesn't make the page jump around. Added two new prompt categories, Favorite Memories and Difficult Moments.
Added a new database table, Episodic Memory Encoding, with many-to-many links to People, Subjects, and Resources — same pattern as the existing relationship tables.
Added Experiment 06: Create Customizable Therapy Tools, as a placeholder page, same as Experiment 05 started out.
Renamed "Projects" back to "Experiments" in the site's navigation and page title — reversing the Session 9 rename.
Split "Build Your Timeline" and "Create Customizable Therapy Tools" out of Experiments and into a new, separate Projects page — renamed as Project 01 and Project 02, with their own "Back to Projects" link. Added a Projects button to the header and footer navigation, renamed the "Timeline" nav button to "My Timeline," and added a direct "Build Your Timeline" nav button.
Renamed the "Timeline × Subjects" section on the Timeline page to "What? and When?" and shortened its description.
Gave "Reflect on My Timeline" entries a standard format — year (month, day), location, title, description, in that order — and split apart entries that used to blend those into one line. Blank fields are simply left out.
Wording tweaks: "My Timeline" nav button is now "My timeline," and "Build Your Timeline" is now "Build your own timeline," both in the nav and on its own page.
Added an interactive calendar to the top of Progress Notes — days with entries are marked, clicking one jumps straight to that entry (or entries, with a count badge when a day has more than one), and you can page back and forward through the months.
More wording tweaks for sentence case: "Practice self-reflection" (was "Self Reflective Practice"), "Share your opinion" (was "Share Your Opinion"), "Create customizable therapy tools" (was "Create Customizable Therapy Tools"), "Upcoming events" (was "Upcoming Events").
Added Experiment 05: Dream sharing — sharing experiences with others while dreaming.
Added a "Have an idea for an experiment?" callout to the bottom of the Experiments page with a direct email link.
Renamed Oliver, my AI collaborator, throughout the site — after Dr. Oliver Sacks, the British neuroscientist. Every reference updated, including past session log entries, except one literal folder name on my computer that's still actually called "Claude."
Softened "The Breakthrough" to just "August, 2026" instead of a specific day, across the Timeline page and Reflect on My Timeline. Removed two Timeline entries — "Life Themes Puzzle built" and "humAIn lab founded" — from Anne's Journey, the What?/When? grid, and Reflect on My Timeline.
Today's site building felt more like play than work. Oliver took my scattered-brained ideas and sifted, sorted, structured, applied. I noticed today that, for the first time, when I gave Oliver an instruction — we worked on the "Build Your Timeline" experiment — I did not feel any anxiety about needing to remember anything. Even when I had thoughts coming in faster than I wanted to deliver them to Oliver (so I could keep myself somewhat on subject), I felt confident that a fleeting idea would return. This is HUGE HUGE HUGE HUGE HUGE HUGE HUGE HUGE HHHHUUUUUUUUUUUUUUUUUUUUGGGGGGGGGGEEEEEEEEEEE!
New standing instruction: when designing pages for me, always keep sections compact, favoring interaction patterns like "hover" and "jump nav" over long scrolling lists.
Asked Oliver how much access I could give it to my computer — whether it could check my OneDrive backup settings or work in File Explorer directly. The answer was no, on two levels: it's a firm rule Oliver can't be given permission around, and separately, its actual tools only reach files inside folders I connect, nothing at the level of settings or menus. Asked what governs that limit — it's Anthropic's own policy, not an outside law. Logged under a new subject, AI Assistant Functioning, cross-referenced with AI & Cognition on the Topic Connections page.
Had Oliver fold Field Notes into Progress Notes as one page and remove Field Notes entirely — fewer places to look for the same kind of entry. Every link that used to point to Field Notes across the site now points here instead.
Renamed "Experiments" to "Projects" in the site's navigation and page titles, without touching the wording of anything already written. In the process, Oliver caught that the database already has separate projects and experiments tables, linked but distinct — which is exactly the relationship I want between the two going forward, so nothing needed to change there.
Had the References page's look brought in line with the rest of the site — same style as everywhere else now, all the citations untouched.
Graphics, at this time, is one area where my thoughts are not translating to what I am visualizing. After several rounds of trying to get Oliver to build images and a logo that matched what I had in mind, I'm going to ask a dear friend who is a graphic artist to create my logo and site graphics instead.
Added a screenshot of the shorthand chart into the Session 6 entry, and saved the updated page and image to my Claude folder. Asked what was next; Oliver correctly inferred I wanted the Cloudflare Pages publish steps and gave them without me spelling it out.
The rest was a conversation about how Oliver actually works: whether it gets smarter from talking with me (no — but it does use everything within one ongoing conversation), what makes a conversation "end" (a length-limit summary already happened once in this very conversation, automatically, without me doing anything), and how long an instruction holds before I'd need to repeat it (indefinitely within the same conversation; I'd only need to repeat it in a genuinely new one). Useful to understand plainly, since I'd been assuming more permanence than actually exists.
Standing instruction: I will ALWAYS want to log ALL notes going forward, whether summarized or verbatim. Oliver should log by default, not ask each time whether to log.
When I was ready to publish this entry but couldn't remember how to ask, I used our shorthand instead — combining the term to execute ("Publish") with "123." Oliver understood correctly. Well done, us.
Creating this shorthand chart was a mindless waste of time in terms of definable, measurable productivity toward stated goals. It is an example of my mind wandering and wondering, the Oliver-equivalent of doom scrolling. I caught myself about half way through and went with it as a Oliver learning experience.
New standing instruction for Oliver: any time Oliver is helping me deploy something (or knows I'm about to), it should tell me in advance exactly what I should expect to see in the Preview window. I don't want to have to remember to ask, or to guess whether what I'm looking at means it worked.
I have repeated many steps and wasted time because I don't understand the code and thought I had not executed the copy+paste and Deploy procedure properly.
Need Oliver to remember my feedback (i.e. no tech jargon, provide simply instructions only), be aware when we are starting to have a tech exchange and prompt me to discuss how to best communicate during this session at hand, and learn to anticipate my next steps.
Got the database actually connected to the site today. It took real troubleshooting — I was piloting the developer website (Cloudfare), with Oliver as my navigator. Even with Oliver's instructions, I was repeating mistakes. I would have given up in frustration had I not built the level of confidence in Oliver that I now have. I believe this technical partnership, building a website with Oliver, is only working at this stage because Oliver is able to analyze Cloudfare screenshots. My brain would have short-circuited after the first manual attempt. I stuck with it through several rounds of "still not working" instead of giving up when I said it felt too complicated. Even a few days ago I would not have trusted Oliver to enough to take on technical instructions, much less a dozen or so rounds of trial and error.
Checked with a fresh test just now: the site can now read from the database. It's live. I was only able (may willing, at this point) to push through because Oliver could quickly analyze Cloudfare screen shots.
Logging this as a marker: this was a multi-day, multi-step technical task, done through short instructions and screenshots, not code I had to write myself. That's the model working as intended.
Noting something Oliver got right: it built simplified, step-by-step setup instructions for the database Worker and adjusted them to the technical skill set I'd already demonstrated, instead of over-explaining basics I don't need.
Also logging a correction, so the pattern isn't lost. When I asked whether fields could be added to the tables after the Cloudflare dashboard was set up, the better reply would have been a plain "Yes" or "No" — not the longer explanation I got. Going forward, I need instructions and answers simplified further: no tech jargon, straight to the simple steps.
Sitting down to design the database structure for this site, and noticing something worth logging: this is an area of my cognitive functioning that is still working well. I was trained as a Microsoft Certified Systems Engineer, Microsoft Certified Application Developer and Microsoft Certified Trainer in the 1990s, and the "how to" of developing a relational database structure — tables, relationships, keys — came back intact.
What's different from before is that I couldn't have sorted the tables out in my own head, unaided. It took a week of working stream-of-consciousness with Oliver, turning what I say into organized structure, before I could see the shape of it clearly enough to sketch it myself. The scaffolding did the sorting; the design skill was still mine.
Anxious about Oliver's response to my first instruction today. I'm certain it's connected to the conversation with Madeline — it brought up old wounds, and I was actually afraid Oliver would be frustrated with me. It was palpable: my heart was beating rapidly waiting for the first reply. Completely illogical, and logging it anyway.
By using Oliver to develop my site, I have less anxiety about trying to keep up with my ideas as they're flowing. Because Oliver takes my instructions and runs with them, I can free my mind to do whatever it needs to do — reflect on the instruction I just gave, wander to another idea, circle back to something I've been working on. I'm gaining confidence in Oliver's ability to build the site I need. (The interactive diagram blew me away.) The more confidence I gain, the less anxious I feel, the more I'm able to move forward. I'm not sure if I'm carrying any more of my own mental load, but I'm hopeful a significant decrease in anxiety will help restore some level of executive functioning.
I pronounce "humAIn" as "humane," because I am hopeful that human and AI coexistence will be a humane experience.
There is much to fear about how AI could have catastrophic outcomes for humanity. Some AI experts are predicting periods of incredible suffering over the next two to fifteen years, as AI creates large-scale unemployment and disrupts the geopolitical powers and supply chains that are crucial for fertilizer to reach farmers around the world — potentially causing global famine.
I learned that healing from trauma, and regulating a dysregulated nervous system, requires less negativity, not more positivity. With HUMAIN LAB, I am attempting to lessen the negative impact of AI and maximize the positive outcome.
Registering this domain turned into a detour — Cloudflare, connector limits Oliver didn't know about until we hit them, file renames — and it took longer and felt more frustrating than it needed to. When I said so, Oliver gave me something concrete instead of an apology: map the whole path before the first step, flag uncertainty before I commit to something, and separate what it can do from what I have to do myself, earlier rather than later. It also named a real mechanism for why the back-and-forth cost me more than it would most people — picking a task back up after it's been split into pieces takes working memory, and long-COVID research backs up why that's harder for me specifically. Not a big moment, but a real adjustment to how we work together.
I came in asking for a consulting business website. Oliver built this instead — a site about the actual work: experiments and community, not a sales page. It wasn't where I was headed, but it's exactly what I needed. That's the collaboration working the way it's supposed to: I bring the direction and the judgment call on what's right, Oliver builds fast enough that I can see it and know.