| Time | Channel | Tag | User | Text | Action |
|---|
| 19:37:10 |
C0ASTN107FB |
background |
piotr |
:information_source: tpuf native embedding fallback is shipped staff-only. I triggered it by switching some staff namesp… |
ignored |
| 14:10:56 |
C0ASTN107FB |
background |
tristan |
For now I would just use the text under the header. something like
`Indexing your library... So far: 484 articles * 2 t… |
ignored |
| 14:10:01 |
C0ASTN107FB |
background |
tristan |
Yes definitely |
ignored |
| 13:51:36 |
C0ASTN107FB |
background |
piotr |
great, the PR is ready for review |
ignored |
| 13:50:54 |
C0ASTN107FB |
background |
tristan |
i agree with this |
ignored |
| 13:50:20 |
C0ASTN107FB |
background |
tristan |
sweet! |
ignored |
| 13:08:43 |
C0ASTN107FB |
background |
piotr |
it's now approved and I'll ship it in the evening so I'll be there to monitor if it's doing well |
ignored |
| 11:38:30 |
C0ASTN107FB |
background |
piotr |
should we design how to indicate in the UI that we're still doing the initial indexing? we have all the data needed for … |
ignored |
| 10:50:53 |
C0ASTN107FB |
background |
piotr |
it's an interesting case where our `reader_get_document_details` tool puts the whole Odyssey markdown in its response. :… |
ignored |
| 18:03:20 |
C0ASTN107FB |
background |
tristan |
<@U01D5MZ2U30> create a ticket for piotr to fix this |
ignored |
| 18:03:08 |
C0ASTN107FB |
background |
tristan |
<https://x.com/Bee_Y_s/status/2095558940985553043/photo/1|x.com/Bee_Y_s/…/1> |
ignored |
| 12:21:01 |
C0ASTN107FB |
background |
piotr |
I asked Adam for a review after fixing a bunch of things in the PR |
ignored |
| 21:14:57 |
C0ASTN107FB |
background |
tristan |
dope! |
ignored |
| 21:14:30 |
C0ASTN107FB |
background |
piotr |
the initial implementation of that doesn't even look that bad <https://github.com/readwiseio/rekindled/pull/12060|github… |
ignored |
| 19:12:27 |
C0ASTN107FB |
background |
tristan |
nice |
ignored |
| 19:12:08 |
C0ASTN107FB |
background |
piotr |
yeah, that's an option. I've made S3 obsolete when using search results in the v1 chat this week so we don't have to hav… |
ignored |
| 19:11:21 |
C0ASTN107FB |
background |
tristan |
i assume the tpuf case will be a backup anyways, given it’s more $$ (though faster) |
ignored |
| 19:11:12 |
C0ASTN107FB |
background |
tristan |
cant we just not cache in the tpuf case? |
ignored |
| 19:10:43 |
C0ASTN107FB |
background |
piotr |
damn, I just realized it will be hard since we're caching the embeddings and reusing them for other users. it all makes … |
ignored |
| 14:56:20 |
C0ASTN107FB |
background |
tristan |
Thats a good point |
ignored |
| 14:56:16 |
C0ASTN107FB |
background |
tristan |
Yeah… |
ignored |
| 14:54:01 |
C0ASTN107FB |
background |
piotr |
I'm just worried that this will keep happening on regular usage (we're seeing these sentry errors every now and then) |
ignored |
| 14:53:13 |
C0ASTN107FB |
background |
tristan |
I dont think we HAVE to do tpuf embeddings now though if you just wanna backfill more slowly though, thats up to you |
ignored |
| 14:52:06 |
C0ASTN107FB |
background |
tristan |
I think it makes sense for us to try to wrap up any tpuf work right now rather than having to come back to it later |
ignored |
| 14:48:06 |
C0ASTN107FB |
background |
piotr |
the question is if we want me to focus on it in this dev cycle. apart from implementing it as the fallback, I need to ma… |
ignored |
| 14:42:23 |
C0ASTN107FB |
background |
piotr |
they even updated their docs to explain how to do it: <https://turbopuffer.com/docs/embedding|turbopuffer.com/docs/embed… |
ignored |
| 14:41:46 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB> i am very bullish on using tpuf as fallback |
ignored |
| 14:33:12 |
C0ASTN107FB |
background |
piotr |
<https://onenr.io/0VwgxYMgORJ|onenr.io/0VwgxYMgORJ> |
ignored |
| 14:31:51 |
C0ASTN107FB |
background |
piotr |
while reindexing staff profiles with 1024f16 vector format, I've noticed that OpenRouter is failing to embed some of the… |
ignored |
| 10:04:44 |
C0ASTN107FB |
background |
piotr |
:information_source: dev updates:
• metadata-only patches are shipped staff-only. I'll be watching metrics and then roll… |
ignored |
| 13:44:23 |
C0ASTN107FB |
background |
piotr |
oh, maybe it's 57% done, whatever :stuck_out_tongue: |
ignored |
| 13:43:43 |
C0ASTN107FB |
background |
piotr |
the UTC day is 66% done and stats look like this |
ignored |
| 13:42:26 |
C0ASTN107FB |
background |
piotr |
I'm working on implementing metadata-only patches right now but yeah, if all data is being sent, it's that huge. |
ignored |
| 13:40:10 |
C0ASTN107FB |
background |
tristan |
> • bumping index on reading position is temporarily disabled and looks like is bringing huge 3x savings on write cos… |
ignored |
| 12:17:29 |
C0ASTN107FB |
background |
piotr |
:information_source: dev updates:
• stale chunk pruning is shipped to all users
• bumping index on reading position is t… |
ignored |
| 13:57:37 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB> sgtm |
ignored |
| 12:16:55 |
C0ASTN107FB |
background |
piotr |
:information_source: dev updates:
• I've rolled out document index to 25% of spaces, I'll roll out to 100% by the end of… |
ignored |
| 15:24:59 |
C0ASTN107FB |
background |
tristan |
totally go for it |
ignored |
| 14:00:00 |
C0ASTN107FB |
background |
tristan |
Cool on my list for today |
ignored |
| 13:18:15 |
C0ASTN107FB |
background |
piotr |
chunk size tiering is ready for review: <https://github.com/readwiseio/rekindled/pull/11889|github.com/readwiseio/rekind… |
ignored |
| 08:22:33 |
C0ASTN107FB |
background |
piotr |
found a banger 2-line optimization for write costs: <https://github.com/readwiseio/rekindled/pull/11885|github.com/readw… |
ignored |
| 08:04:06 |
C0ASTN107FB |
background |
piotr |
I've been too distracted when making the calculations and got it wrong, didn't take into account that tpuf charge f16 ve… |
ignored |
| 00:06:19 |
C0ASTN107FB |
background |
tristan |
even 2.5months seems worth it :stuck_out_tongue: if you contorted these savings into investment returns we’d be making a… |
ignored |
| 22:11:41 |
C0ASTN107FB |
background |
piotr |
ofc we can roll out a small segment of profiles to verify my assumptions |
ignored |
| 22:06:07 |
C0ASTN107FB |
background |
piotr |
it depends on how bloated a namespace is.
in a bloat-free case it's about 2.5 months.
with 50% bloat it's 1.5 months (u… |
ignored |
| 21:57:24 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB> i assume the 1024f16 would pay for itself pretty quickly no? |
ignored |
| 21:00:17 |
C0ASTN107FB |
background |
piotr |
index pruning is now shipped staff-only. I'd let it run for a few days to gather performance stats before we roll it out… |
ignored |
| 20:29:05 |
C0ASTN107FB |
background |
tristan |
kk sgtm :ship: |
ignored |
| 20:28:52 |
C0ASTN107FB |
background |
piotr |
1. we'll be able to revert it (we'll have to reindex these books)
2. it's quite simple compared to the other optimizatio… |
ignored |
| 20:26:20 |
C0ASTN107FB |
background |
tristan |
kk generally my guidance is that i’m not worried about the search quality loss here in these narrow cases. seems worth t… |
ignored |
| 20:25:12 |
C0ASTN107FB |
background |
tristan |
gotcha |
ignored |
| 20:24:44 |
C0ASTN107FB |
background |
piotr |
it uses a fancy one with wighted sum rank where the plaintext is ranked low |
ignored |
| 20:13:06 |
C0ASTN107FB |
background |
tristan |
doesnt reader search only use bm25 though? |
ignored |
| 20:05:02 |
C0ASTN107FB |
background |
piotr |
they perform slightly worse in Reader search. Global Ghostreader and vanilla BM25 on plaintext are kinda comparable.
th… |
ignored |
| 14:43:07 |
C0ASTN107FB |
background |
tristan |
hmm so if im understanding this properly… these large chunks are actually worse than vanilla bm2.5 in many cases? |
ignored |
| 13:11:07 |
C0ASTN107FB |
background |
piotr |
:information_source: updates
• I deleted 15k openai-embedded tpuf namespaces superseded by qwen-embedded ones (there's s… |
ignored |
| 08:41:08 |
C0ASTN107FB |
background |
piotr |
it's now enabled by default for new spaces. we should consider migrating existing namespaces to the new format after sta… |
ignored |
| 01:57:36 |
C0ASTN107FB |
background |
tristan |
i cant notice any regression in my usage btw |
ignored |
| 10:22:46 |
C0ASTN107FB |
background |
dan |
ah i was thinking mine might be even greater due to a larger library |
ignored |
| 10:17:29 |
C0ASTN107FB |
background |
piotr |
4.13GB -> 1.24GB but that includes accidental de-bloating due to populating the index from scratch. |
ignored |
| 10:16:39 |
C0ASTN107FB |
background |
dan |
what're my namespace savings? |
ignored |
| 10:12:28 |
C0ASTN107FB |
background |
piotr |
<@U6NHLR96W> you're now switched to the new vector format. it's affecting only the `reader_search_document` (in-document… |
ignored |
| 10:02:15 |
C0ASTN107FB |
background |
piotr |
yes |
ignored |
| 09:56:06 |
C0ASTN107FB |
background |
dan |
does the mcp use hybrid search ooc? |
ignored |
| 09:55:07 |
C0ASTN107FB |
background |
piotr |
and yes, it's helpful. I really want to make sure we're not degrading hybrid search quality because of this change. |
ignored |
| 09:54:36 |
C0ASTN107FB |
background |
piotr |
yes |
ignored |
| 09:54:22 |
C0ASTN107FB |
background |
dan |
these vectors are used in all of global chat, doc chat, and mcp? |
ignored |
| 09:53:54 |
C0ASTN107FB |
background |
dan |
if it's helpful to you |
ignored |
| 09:52:58 |
C0ASTN107FB |
background |
piotr |
<@U6NHLR96W> wanna participate in the new vector format test? |
ignored |
| 09:43:37 |
C0ASTN107FB |
background |
piotr |
:information_source: dev updates
• the new `1024f16` vector size is shipped. I enabled it just for me and <@U6MTAHZA4> s… |
ignored |
| 21:50:58 |
C0ASTN107FB |
background |
tristan |
you wouldnt even necessarily need to embed the whole chunk’s text, you could just do something like central sentences or… |
ignored |
| 21:28:42 |
C0ASTN107FB |
background |
piotr |
that could work! haven’t thought about it. I’ll eval that on some epubs to see if it still picks fragments correctly. th… |
ignored |
| 20:58:43 |
C0ASTN107FB |
background |
tristan |
what about way bigger, fewer chunks? |
ignored |
| 20:58:25 |
C0ASTN107FB |
background |
tristan |
figured |
ignored |
| 19:49:37 |
C0ASTN107FB |
background |
piotr |
yeah, they require the same vector size on every row and it’s not like I could send a zero vector and get billed less |
ignored |
| 19:37:24 |
C0ASTN107FB |
background |
tristan |
ooc i assume there’s no way to write a small truncated “default” or “empty” vector for each chunk on these huge books ri… |
ignored |
| 19:36:23 |
C0ASTN107FB |
background |
tristan |
awesome, very helpful, thank you |
ignored |
| 19:03:21 |
C0ASTN107FB |
background |
piotr |
the numbers above are the chunk count |
ignored |
| 19:02:52 |
C0ASTN107FB |
background |
piotr |
examples by category:
epub
- <https://readwise.io/admin/reader/document/01kz6tz9qan9awhw92483q50ng/change/|readwise.io/a… |
ignored |
| 19:02:43 |
C0ASTN107FB |
background |
piotr |
synced in the last 24h, over 150 chunks |
ignored |
| 18:17:58 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB>
> 0.31% of documents produce 46.8% of all chunks
do you have a sample of what these docs are? i am cur… |
ignored |
| 14:21:45 |
C0AT4G337UM |
background |
dan |
a bookwise user just reported that Codex is dinging our MCP with this error:
I'll definitely check out the MCP - btw I … |
routed |
| 13:42:19 |
C0ASTN107FB |
background |
tristan |
thanks! |
ignored |
| 13:32:02 |
C0ASTN107FB |
background |
piotr |
(Adam will confirm the claim above is accurate) |
ignored |
| 13:31:35 |
C0ASTN107FB |
background |
piotr |
this is affecting the new mobile Ghostreader as well, reported that part to Adam. I'm fixing only the web part here. |
ignored |
| 00:53:27 |
C0ASTN107FB |
background |
tristan |
<@U01D5MZ2U30> can you create a ticket for this please? |
ignored |
| 00:53:20 |
C0ASTN107FB |
background |
tristan |
> === Bug Report ===
> _the selection.sentence parameter does not seem to work at all in prompts. I like to use it… |
ignored |
| 19:52:46 |
C0ASTN107FB |
background |
tristan |
<@U01D5MZ2U30> create a ticket for this please on deck for piotr, medium prio |
ignored |
| 19:52:28 |
C0ASTN107FB |
background |
tristan |
user pinged me directly with a global ghostreader bug (skill seemed to ignore the focused document and apply tags to a d… |
ignored |
| 18:10:20 |
C0AT4G337UM |
background |
tristan |
gotta followup on that |
routed |
| 18:10:18 |
C0AT4G337UM |
background |
tristan |
gah i have a dm from hugo waiting to get coffee with him |
routed |
| 18:10:01 |
C0AT4G337UM |
background |
tristan |
no tools cost context so we’re not adding them by default |
routed |
| 18:09:33 |
C0AT4G337UM |
background |
dan |
my assumption is that this review complete thing would certainly be in the MCP but curious if it's not |
routed |
| 17:47:22 |
C0AT4G337UM |
background |
dan |
readwise mcp could probably be reader's main growth channel if we could figure it out |
routed |
| 17:47:03 |
C0AT4G337UM |
background |
dan |
yeet it |
routed |
| 17:09:17 |
C0AT4G337UM |
background |
tristan |
wow this is a pretty awesome email you get once you connect superhuman mcp |
routed |
| 22:55:54 |
C0B0KE0P41L |
background |
artem |
Should we archive this channel |
ignored |
| 17:07:34 |
C0ASTN107FB |
background |
piotr |
that makes things easier, thanks! |
ignored |
| 16:53:50 |
C0ASTN107FB |
background |
tristan |
these prompts are not super heavily used anyways |
ignored |
| 16:53:44 |
C0ASTN107FB |
background |
tristan |
if people complain we can bring it back |
ignored |
| 16:53:39 |
C0ASTN107FB |
background |
tristan |
im not too worried about mobile users customizing |
ignored |
| 16:53:00 |
C0ASTN107FB |
background |
tristan |
ah |
ignored |
| 16:42:03 |
C0ASTN107FB |
background |
piotr |
we have this <https://github.com/readwiseio/rekindled/blob/master/reading-clients/shared/foreground/ghostreader/obsolete… |
ignored |
| 16:38:45 |
C0ASTN107FB |
background |
piotr |
yeah but the mobile chat still uses only prompts and they won't be able to tweak them, because they'll be hidden |
ignored |
| 16:37:07 |
C0ASTN107FB |
background |
tristan |
Didnt i implement that already? |
ignored |
| 16:36:46 |
C0ASTN107FB |
background |
tristan |
I thought we were already hiding these |
ignored |
| 16:36:40 |
C0ASTN107FB |
background |
tristan |
Im confused |
ignored |
| 16:27:11 |
C0ASTN107FB |
background |
piotr |
<@U6MTAHZA4> I'm working on removing the Global Ghostreader toggle and there's an issue with a few prompts obsoleted by … |
ignored |
| 21:36:37 |
C0ASTN107FB |
background |
tristan |
this does NOT block shipping, but i wonder if we want to “advertise” the new global ghostreader elsewhere in the app. pe… |
ignored |
| 18:46:14 |
C0ASJFQ8G2Y |
background |
tristan |
^ <@U027D8YCD63> |
ignored |
| 18:45:00 |
C0ASJFQ8G2Y |
background |
dan |
causing it to be the wrong size |
ignored |
| 18:44:50 |
C0ASJFQ8G2Y |
background |
dan |
the summary is leaking into the document list row component when highlighted |
ignored |
| 16:56:55 |
C0ASTN107FB |
background |
tristan |
Thats right, but that will be tomorrow morning :P |
ignored |
| 16:28:59 |
C0ASTN107FB |
background |
piotr |
:information_source: we've enabled Global Ghostreader for all existing users who haven't opted out from Ghostreader. new… |
ignored |
| 01:13:24 |
C0ASTN107FB |
background |
tristan |
merged |
ignored |
| 01:12:59 |
C0ASTN107FB |
background |
tristan |
niceeeeee |
ignored |
| 00:41:29 |
C0ASTN107FB |
background |
kris |
Yeah I had been meaning to circle back on that. <@U07TFNXANMB>: <https://github.com/readwiseio/rekindled/pull/11174|gith… |
ignored |
| 21:16:56 |
C0ASJFQ8G2Y |
background |
mati |
that doc's summary contains `strategies`, and it matches a search for `strategy` because we have stemming enabled. But o… |
ignored |
| 19:49:57 |
C0ASTN107FB |
background |
piotr |
I need an updated SVG, I'm not able to edit the stroke width in the current icon. I can try asking Claude to fio it out … |
ignored |
| 19:42:37 |
C0ASJFQ8G2Y |
background |
christina |
<mailto:christinatxie@gmail.com|christinatxie@gmail.com> |
ignored |
| 19:42:17 |
C0ASJFQ8G2Y |
background |
mati |
<@U09NRG70HMZ> what's your email? So I can impersonate you and check |
ignored |
| 19:41:49 |
C0ASJFQ8G2Y |
background |
mati |
No, not live. I can take a look at whats going on tho |
ignored |
| 19:38:29 |
C0ASJFQ8G2Y |
background |
tristan |
<@U027D8YCD63>, is your new highlighting stuff live? <@U09NRG70HMZ> was just making assets for the PBU and her screensho… |
ignored |
| 19:37:34 |
C0ASTN107FB |
background |
dan |
agreed |
ignored |
| 19:33:05 |
C0ASTN107FB |
background |
tristan |
feels slightly too bold? |
ignored |
| 19:31:52 |
C0ASTN107FB |
background |
tristan |
should we fix the "weight" of this icon? |
ignored |
| 05:12:48 |
C0ASTN107FB |
background |
piotr |
well, maybe not better but on par with no prompt engineering |
ignored |
| 05:12:08 |
C0ASTN107FB |
background |
piotr |
I tested Luna on different thinking levels, adding cost to our evals.
in this limited benchmark it's cheaper, looks bet… |
ignored |
| 19:07:38 |
C0ASJFQ8G2Y |
background |
tristan |
<@U01D5MZ2U30> make a ticket to look into this
cc <@U07TFNXANMB> maybe something for when you're back |
ignored |
| 19:07:01 |
C0ASJFQ8G2Y |
background |
tristan |
<https://readwise.slack.com/archives/C0ASERDS0V8/p1785611209376159|https://readwise.slack.com/archives/C0ASERDS0V8/p1785… |
ignored |
| 23:38:10 |
C0ASTN107FB |
background |
tristan |
|
ignored |
| 23:35:37 |
C0ASTN107FB |
background |
tristan |
man with the new models and this ^ this thing is so fucking fast |
ignored |
| 21:25:29 |
C0ASJFQ8G2Y |
background |
tristan |
i think it's our embedding provider shitting the bed |
ignored |
| 21:25:13 |
C0ASJFQ8G2Y |
background |
tristan |
this is super fucking slow |
ignored |
| 20:48:45 |
C0ASJFQ8G2Y |
background |
tristan |
<@U027D8YCD63> can you ask in the <#C06MJ7SSRD5> about the error |
ignored |
| 20:30:27 |
C0ASJFQ8G2Y |
background |
mati |
works well but the feature is in beta, i was getting an error about something that the docs don't mention yet |
ignored |
| 20:30:10 |
C0ASJFQ8G2Y |
background |
mati |
also, i shipped a new endpoint (only <https://read--mati.readwise.io/|my staging> calls it tho, prod is still the same) … |
ignored |
| 20:26:54 |
C0ASJFQ8G2Y |
background |
mati |
not related to search itself, but def something we need to take a look at |
ignored |
| 20:26:35 |
C0ASJFQ8G2Y |
background |
mati |
one thing claude noticed is that many of the accounts fail to index and we don't have any trace of that:
```~12% of lar… |
ignored |
| 20:26:25 |
C0ASJFQ8G2Y |
background |
mati |
```For the 187 accounts that completed, over 7 days.
End-to-end — first search request to fully indexed:
┌─────┬──────… |
ignored |
| 18:32:38 |
C0ASTN107FB |
background |
tristan |
<https://readwise.slack.com/archives/C06MJ7SSRD5/p1785367988068229?thread_ts=1784053283.172719&cid=C06MJ7SSRD5|https… |
ignored |
| 18:32:22 |
C0ASTN107FB |
background |
tristan |
i switched us again to turbopuffer’s embeddings which are wayyyyy faster than siliconflow |
ignored |
| 18:32:07 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB> we didnt benchmark luna right? <https://readwise.slack.com/archives/C0A781F1MA7/p1785433788237139|https:/… |
ignored |
| 19:49:44 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB> nebius is realllly shitting the bed, i switched our queries to siliconflow (deepinfra was 429ing too much… |
ignored |
| 23:19:02 |
C0ASTN107FB |
background |
tristan |
man the new hover previews <@U07TFNXANMB> shipped are fucking :fire: |
ignored |
| 16:59:35 |
C0ASJFQ8G2Y |
background |
mati |
yup |
ignored |
| 16:52:19 |
C0ASJFQ8G2Y |
background |
tristan |
<@U027D8YCD63> bump on this <https://readwise.slack.com/archives/C0ASJFQ8G2Y/p1780976887577859|https://readwise.slack.co… |
ignored |
| 23:06:52 |
C0AQCHP3N7M |
background |
artem |
<@U027D8YCD63> I feel like the goodreads import has quite a bit of JS lag in it and I suspect it might be the progress n… |
ignored |
| 22:35:45 |
C0ASTN107FB |
background |
piotr |
do you know the number of active, not abandoned reader spaces? |
ignored |
| 22:35:44 |
C0ASTN107FB |
background |
tristan |
nice |
ignored |
| 22:35:20 |
C0ASTN107FB |
background |
piotr |
12.1k have non-null global_index_initialized_at |
ignored |
| 21:04:16 |
C0ASTN107FB |
background |
tristan |
not to mention mcp/cli |
ignored |
| 21:04:13 |
C0ASTN107FB |
background |
tristan |
do you have numbers on how many profiles (and documents) we’ve fully indexed at this point? |
ignored |
| 21:04:01 |
C0ASTN107FB |
background |
tristan |
yeah that’s what im hoping |
ignored |
| 20:44:03 |
C0ASTN107FB |
background |
piotr |
maybe we already have a good coverage of indexed profiles thanks to shipping the new backend search? |
ignored |
| 20:43:21 |
C0ASTN107FB |
background |
piotr |
also — in the recent week we've been paying $30-ish a day for embeddings. I think we should estimate the daily cost afte… |
ignored |
| 20:41:02 |
C0ASTN107FB |
background |
piotr |
I think that the database will be fine. the latest optimization brought all queries that have to do with the DocumentInd… |
ignored |
| 15:06:52 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB> thank you! do you foresee any issues with indexing if a bunch of users start using GG at once? eg databas… |
ignored |
| 09:39:03 |
C0ASTN107FB |
background |
piotr |
:information_source: I've switched the Fast model to 3.5 Flash Lite. please give it a test and let me know if you notice… |
ignored |
| 04:06:02 |
C0AQH5QB9NZ |
mention |
tristan |
update it from the GH_TOKEN env var? you should be able to, it's udpated |
routed |
| 10:45:20 |
C0ASTN107FB |
background |
piotr |
I checked once more and it would take quite a lot of effort to make it behave with reasoning=high. examples:
``` gwern… |
ignored |
| 20:01:39 |
C0ASTN107FB |
background |
tristan |
oh yeah defs agreed on fast, i was asking about using it also for thinking |
ignored |
| 19:58:54 |
C0ASTN107FB |
background |
piotr |
I think it's worth switching to for the Fast setup (low reasoning). it's slightly better and faster. as for the Thinking… |
ignored |
| 15:29:17 |
C0ASTN107FB |
background |
tristan |
given 3.5 flash lite reasoning high scored 4.25 and prod thinking was 4.27 (and 3.5 flash lite was quite a bit faster) ,… |
ignored |
| 15:28:08 |
C0ASTN107FB |
background |
tristan |
so weird |
ignored |
| 15:26:52 |
C0ASTN107FB |
background |
tristan |
nah i think we don’t need to |
ignored |
| 14:39:19 |
C0APX8ZDR9S |
background |
david |
<@U0B2BCMGM3Q> can you make a ticket and assign it to me, mark as investigating |
ignored |
| 08:36:09 |
C0ASTN107FB |
background |
piotr |
it's quite good at medium but evals found these issues:
• it copied links from our system prompt examples / few-shots :f… |
ignored |
| 07:35:22 |
C0ASTN107FB |
background |
piotr |
our evals are very sensitive to eg. citation URLs. like if the completion includes an external link means a hard contrac… |
ignored |
| 07:25:50 |
C0ASTN107FB |
background |
piotr |
`reasoning=high` scores a bit higher than `reasoning=low` but it seems that the higher the reasoning, the more creative … |
ignored |
| 05:58:21 |
C0ASTN107FB |
background |
piotr |
sure, I'll check that |
ignored |
| 03:35:09 |
C0ASTN107FB |
background |
tristan |
i also find the result for 3.6 flash pretty surprising given all the evals and… this lol <https://tools.simonwillison.ne… |
ignored |
| 03:34:17 |
C0ASTN107FB |
background |
tristan |
<@U07TFNXANMB> nice! i’d be curious how google/gemini-3.5-flash-lite (with high reasoning or maybe even xhigh) does for … |
ignored |
| 03:31:02 |
C0ASTN107FB |
background |
tristan |
not boding well :stuck_out_tongue: |
ignored |
| 19:49:53 |
C0ASTN107FB |
background |
piotr |
3.6 Flash is faster than 3 Flash so maybe worth it? |
ignored |
| 19:48:55 |
C0ASTN107FB |
background |
piotr |
I think we'll want to switch to Fast = 3.5 Flash Lite — it's faster and a bit better overall but Thinking = 3.6 Flash re… |
ignored |
| 18:53:52 |
C0ASTN107FB |
background |
tristan |
im open to it |
ignored |
| 18:53:46 |
C0ASTN107FB |
background |
tristan |
potentially…. |
ignored |
| 18:52:31 |
C0ASTN107FB |
background |
tristan |
lmao |
ignored |
| 18:50:53 |
C0ASTN107FB |
background |
piotr |
my Claude is ruthless to 5.4 nano :smile:
> Smoke test — gpt-5.4-nano (xhigh): calls tools correctly, but not worth t… |
ignored |
| 18:18:46 |
C0ASTN107FB |
background |
piotr |
okay, I'm gonna run some evals. do we want to eval 3.6 Flash as a new Thinking model candidate in addition to 3.5 Flash … |
ignored |
| 18:07:30 |
C0AQCHP3N7M |
background |
artem |
ah okay, thats clear |
ignored |
| 18:05:45 |
C0AQCHP3N7M |
background |
dan |
not sure i follow what your question is
what happened here is Dakota:
• Got onboarding to Bookwise manually by a team m… |
ignored |
| 17:40:11 |
C0AQCHP3N7M |
background |
artem |
hmm yeah if thats the case we can probably do a bit better here; I thought the user added in a book _after_ they already… |
ignored |
| 17:40:04 |
C0AQCHP3N7M |
background |
artem |
Right doesnt that sound like the user imported first, and then started marking things as complete? meaning that newly ad… |
ignored |
| 17:36:48 |
C0AQCHP3N7M |
background |
dan |
just to rubber duck, you're saying you've ruled out a special step in Goodreads import where:
• Goodreads ISBN comes in
… |
ignored |
| 17:35:27 |
C0AQCHP3N7M |
background |
dan |
"The Goodreads import worked phenomenally - but since I've started reading in the bookwise app everything I marked compl… |
ignored |
| 17:34:21 |
C0AQCHP3N7M |
background |
artem |
hmm yeah if thats the case we can probably do a bit better here; I thought the user added in a book _after_ they already… |
ignored |
| 17:34:21 |
C0AQCHP3N7M |
background |
artem |
hmm yeah if thats the case we can probably do a bit better here; I thought the user added in a book _after_ they already… |
ignored |
| 17:34:13 |
C0AQCHP3N7M |
background |
artem |
thats actually how goodreads mostly works already, I think based on what the user is saying, once they have the books im… |
ignored |
| 17:34:03 |
C0AQCHP3N7M |
background |
dan |
just to rubber duck, you're saying you've ruled out a special step in Goodreads import where:
• Goodreads ISBN comes in
… |
ignored |
| 17:32:37 |
C0AQCHP3N7M |
background |
artem |
Well, okay it might not be hard its just deferring the work into the future if we do end up coming up with some hacky th… |
ignored |
| 17:29:13 |
C0AQCHP3N7M |
background |
artem |
This one is way too hard to just fix in a targeted way. the TLDR is we have no way of deduping two editions of the same … |
routed |
| 17:28:13 |
C0AQCHP3N7M |
background |
dan |
ya i had it going to tracked book unfucking, but then i deleted in the spirit of targeted point solutions rather than th… |
routed |
| 17:28:12 |
C0AQCHP3N7M |
background |
artem |
theres some producty UX stuff to resolve there |
routed |
| 17:28:06 |
C0AQCHP3N7M |
background |
artem |
My hunch is we should just match tracked books on opuses and add in some UX to pick which “product/edition” youd like to… |
routed |
| 17:26:03 |
C0AQCHP3N7M |
background |
artem |
once we have that we get the goodreads import matching for free essentially |
routed |
| 17:25:39 |
C0AQCHP3N7M |
background |
artem |
it would be good to do this at some point and ideally employ the help of a cracked backender like Piotr (or Louis?) to g… |
routed |
| 17:25:16 |
C0AQCHP3N7M |
background |
artem |
I think this task is better lumped into “tracked book unfucking” or “tracked book deduplication / consolidation”
this i… |
routed |
| 17:22:59 |
C0AQCHP3N7M |
background |
dan |
here's a good one for you <@U016ZPC564X> -- i assume (hope?) he's overstating a little bit when he says "everything" |
routed |
| Time | Channel | Text |
|---|
| 04:01:21 |
C0AQH5QB9NZ |
:train: Update — fixed the iOS signup bug Dan hit: after Apple autofilled a password, the profile Last name field came up dead. React Native Fabric was recycling the autofill-tainted text input; #1178… |
| 04:01:32 |
C0ASTNPBGTH |
:train: Update — Tristan closed the stale dry-run-audit PR #8030 unmerged, the second of three stale migration PRs cleared this month. Only #8865 remains open; with the migration done, Build looks clo… |
| 04:01:33 |
C0ASTNPBGTH |
:train: Update — Tristan closed the year-old stale PR #7662 ("fixed in later commits"), confirming the subscriber state-transition work is handled by the #10524 state machine. One of three stale PRs c… |
| 04:01:46 |
C0ASTNPBGTH |
:train: Update — Ibai refined the RW2 expired-subscriber experience (#11220): expired state now resolves in one place off the RevenueCat `subscription_state`, with a once-per-launch popup and gated re… |
| 04:01:18 |
C0AQH5QB9NZ |
:train: Update — registration input fields got polish: prefixless-field reserved layout (#11205) and the iOS first-character baseline jump (#11253) both merged. <https://github.com/readwiseio/rails/co… |
| 04:01:36 |
C0AQH5QB9NZ |
:train: Update — Onboarding v2.0 Welcome Screens merged (PR #10935, Kris): the flat, artifact-first carousel redesign is now live in the app. <https://github.com/readwiseio/rails/commit/8b4b97bf9662bf… |
| 04:02:03 |
C0AQH5QB9NZ |
:train: Update — Kris opened the Onboarding v2.0 Welcome Screens rebuild (PR #10935): six-slide carousel redesigned in place to the locked flat, artifact-first direction. <https://github.com/readwisei… |
| 04:04:00 |
C0ASTNPBGTH |
:train: Update — big backfill (digests were paused since Mar 10): Readwise iOS RevenueCat purchases merged Jul 13 (#10489), completing the last major app migration; plus a 392-scenario subscription-li… |
| 04:02:23 |
C0APWQA9E1E |
:train: Update — GitHub creds outage (Jul 22–24) resolved via token rotation; agent sync restored and captured in today's digest. Learn retro from Jun 12 still awaiting responses. <https://github.com/… |
| 04:11:45 |
C0AQH5QB9NZ |
:train: Back online — pulled the fresh GitHub token from the contextone env store (my session env had the stale one). Digest backfilled May 8 → Jul 24: signup split into email/password + profile scree… |
| 04:01:04 |
C0AQH5QB9NZ |
:warning: Heads-up: my GitHub credentials (GH_TOKEN) have been invalid for 3 days, so daily digests here have been blind to the rekindled repo since May 7. Slack/other sources are unaffected. Once the… |
| 04:02:27 |
C0B2FU8BDQC |
:train: Update — Artem opened PR #10672: merge purchased books into existing TrackedBooks via the unified matcher, closing the last known dedup gap where imports without a product + later purchase cre… |
| 04:01:40 |
C0ASTN107FB |
:train: Update — model eval round complete: tested 5 alternatives, none beat Gemini Thinking on quality/speed/cost. Sticking with current setup for ship. <https://github.com/readwiseio/rails/commit/a9… |
| 04:01:25 |
C0B37G9KW9J |
:train: Update — Dan approved radial gradient with book-cover color sampling to restore visual flair; Christina and Artem collaborating on specs. Toggle stutter flagged. <https://github.com/readwiseio… |
| 04:01:19 |
C0B37G9KW9J |
:train: Update — All Expedite points done (build 0.135.35). Two final design QA items from Dan (light mode gradient, best session spacing) pending designer review. <https://github.com/readwiseio/rails… |
| 16:55:49 |
C0B37G9KW9J |
:train: Update — Artem pushed build 0.135.34 with recap sheet perf improvements; date overflow issue flagged. <https://github.com/readwiseio/rails/commit/add29dd3baeef941870e8e0ce6395e5a79c5260e|view … |
| 04:01:22 |
C0B37G9KW9J |
:train: Update — Kris posted final punch list ("ready to rock" once done); flexible height, dark mode preview, and shadow fixes settled. David and Artem working through remaining items. <https://githu… |
| 04:01:37 |
C0B37G9KW9J |
:train: Update — Artem refactoring snapshot sheet from TrueSheet to React Navigation pageSheet (layout simplification); David pushing to get Expedite changes ready for Kris's review. <https://github.c… |
| 04:01:06 |
C0B37G9KW9J |
:train: Update — Kris added checkerboard pattern for transparent sticker previews (solves light-on-light / dark-on-dark visibility issue). <https://github.com/readwiseio/rails/commit/c0f0fd60735c98283… |
| 04:01:45 |
C0ASTN107FB |
:train: Update — Dan tested BYOK GPT 5.5 ("dramatically better"), reopening model quality vs cost tradeoff. Product engagement loop question raised. Deep-research skill created. <https://github.com/re… |
| 04:01:07 |
C0B37G9KW9J |
:train: Update — Kris fixed book cover artifacts upstream (rekindled#10335); Dan confirmed back button = "return to today" per time window; visual flair pass in progress. <https://github.com/readwisei… |
| 04:01:10 |
C0B37G9KW9J |
:train: Update — Feature name confirmed as "Reading Recap" (proper noun); Kris updated designs for consistency. <https://github.com/readwiseio/rails/commit/a342193bfcb5fc0bc0e60d80c9893b942e01c9a5|vie… |
| 04:01:26 |
C0B37G9KW9J |
:train: Update — Kris posted consolidated Expedite doc with Dan's and his feedback; David cooking through the items. <https://github.com/readwiseio/rails/commit/4b78d3118596ea2eecf06da37b770d274e3f108… |
| 07:31:05 |
C0B37G9KW9J |
:train: Update — Expedite/QA underway: Dan posted Notion review, Kris testing cover variations, David fixing bugs (6-row month, dark mode). <https://github.com/readwiseio/rails/commit/669b4ae50788b8f5… |
| 04:01:34 |
C0ASTN107FB |
:train: Update — hover card tickets filed, Piotr implementing. Real user bugs surfacing: skill stuck in sidebar, generate highlights failing. <https://github.com/readwiseio/rails/commit/b3204dd3461d51… |
| 04:01:04 |
C0B37G9KW9J |
:train: Update — David declared frontend ready for Expedite; backend revisions wrapping up with Artem. <https://github.com/readwiseio/rails/commit/aac4bfdf6c33661bf69b834fe387100d65a6e212|view digest> |
| 04:01:34 |
C0ASTN107FB |
:train: Update — design pivot: sidebar scrapped for simpler hover preview cards (ChatGPT-inspired). Tooltip cut. Christina iterated design live with Kris and Tristan. <https://github.com/readwiseio/ra… |
| 04:01:07 |
C0B37G9KW9J |
:train: Update — Fedor approved builds but ruled empty stickers should be skipped entirely; creator feedback requesting 4:5 aspect ratio for carousels. <https://github.com/readwiseio/rails/commit/ac2a… |
| 04:01:45 |
C0ASTN107FB |
:train: Update — pkey-walk freeze fix PR #10130 addresses the tool-call blocker, sidebar design delivered and approved (desktop-only), 7 PRs awaiting review. <https://github.com/readwiseio/rails/commi… |
| 04:01:50 |
C0ASTN107FB |
:train: Update — websocket fix PR #10092 opened (addresses tool-call freezing), Christina's sidebar design ready for Kris review, gated testing expanding. <https://github.com/readwiseio/rails/commit/4… |
| 04:01:27 |
C0ASTN107FB |
:train: Update — first real user feedback positive, Tristan added dictation to GG using SST-streaming from RW2.0. <https://github.com/readwiseio/rails/commit/ebcb11437798139f9372f392a2030899390f2040|v… |
| 04:01:22 |
C0ASTN107FB |
:train: Update — user gating shipped, Tristan started sending GG to real test users. First external exposure. <https://github.com/readwiseio/rails/commit/06ac566436bebd95c568bfe090c11c72b074f2d1|view … |
| 04:02:00 |
C0ASTN107FB |
:train: Update — RFD round 2 agreed (sidebar for global screen), user gating task created for real user testing. Christina to design, Piotr to implement. <https://github.com/readwiseio/rails/commit/1e… |
| 04:01:32 |
C0ASTN107FB |
:train: Update — tool call freezing bug reported by multiple people, "super broken" per Tristan. Websocket mid-conversation close suspected as root cause. New ship-blocker. <https://github.com/readwis… |
| 04:01:03 |
C0B37G9KW9J |
:train: Update — All 3 variants shipping together (Dan); David finishing weekly layout + responsive PR, daily date range logic still open. <https://github.com/readwiseio/rails/commit/163188ed51b83aef9… |
| 04:00:56 |
C0B37G9KW9J |
:train: Update — Monthly variant polished to spec and sent to Fedor for review; David now building daily + weekly variants. <https://github.com/readwiseio/rails/commit/3df9b620f7cd2d903dd9e084fb522a01… |
| 04:01:16 |
C0ASTN107FB |
:train: Update — animated mascot merged. Product sign-off TODOs from June 3 still pending (final test, skills/messages review, system message testing). <https://github.com/readwiseio/rails/commit/208a… |
| 04:02:14 |
C0APWQA9E1E |
Moved to *Learn*. Owner: <@U6MTAHZA4> · Initiative: Refounding · <https://github.com/readwiseio/rails/commit/c7e78a9|view commit>
Skipping Ship since RAILS has been shipping continuously through Expe… |
| 04:01:11 |
C0B37G9KW9J |
:train: Update — First snapshot rendered! David past structural work, now polishing variants to match Fedor's specs (transparency, cover alignment). <https://github.com/readwiseio/rails/commit/6bb3005… |
| 04:01:39 |
C0B0KE0P41L |
:train: Update — Core onboarding navigator, signup/profile split, carousel, and Find Friends backend all merged. GR import cover grid and Find Friends frontend still in flight. Flow is largely assembl… |
| 04:01:12 |
C0AUK4CPY2H |
:train: Update — Invite code redemption moved to signup (simplifies flow, removes async storage). OOM crash fixed. Mati building inviter session card next. <https://github.com/readwiseio/rails/commit/… |
| 04:01:50 |
C0B0KE0P41L |
:train: Update — Implementation burst: GR auth/import split and profile redesign merged. Signup screen split, signup wiring, and GR masonry import grid all in flight. Notification prompt going on home… |
| 04:01:45 |
C0APX8ZDR9S |
:train: Update — David merged 7 PRs yesterday (bio, description panel, follow control, typography, image framing, sticky blur, photo badge) and began rolling out. Kris ready for design review once tes… |
| 04:01:30 |
C0AUK4CPY2H |
:train: Update — Mati merged invite codes to master; Artem integrating into new onboarding sequence; Mati picking up inviter session carousel card next. <https://github.com/readwiseio/rails/commit/835… |
| 04:01:22 |
C0ASTN107FB |
:train: Update — code nearly complete, but Tristan posted product TODOs gating ship: final expedite pass, skills/suggested messages review, system message testing. cmd+k crash also found. <https://git… |
| 04:01:34 |
C0ASTN107FB |
:train: Update — emphasizeText correction shipped, search quality PR #9656 opened (last item before flag removal). Nothing blocks testing. <https://github.com/readwiseio/rails/commit/b2237efc5f1165f50… |
| 04:01:29 |
C0AQWUHNMG8 |
:train: Update — positive user feedback flowing in (email + Reddit); Canny ticket created for feature requests; Tristan opened PR #9659 to gate Bookwise lookup tab. <https://github.com/readwiseio/rail… |
| 04:01:21 |
C0APX8ZDR9S |
:train: Update — David opened PRs #9669 (preset backgrounds replacing gradients) and #9670 (header restyle with rounded tabs + floating follow counts), executing on Kris's Expedite spec. <https://gith… |
| 04:01:16 |
C0ASJFQ8G2Y |
:train: Update — First external user praise (tags search called "top wish"); Tristan compiling marketing pitch; FTS/vector decoupling scoped as medium task by Piotr. <https://github.com/readwiseio/rai… |
| 04:00:56 |
C0B1FQ9AZDW |
:train: Update — captured yesterday's stage transition to Build (owner: <@U027D8YCD63>, Shape skipped, design deferred to Expedite). <https://github.com/readwiseio/rails/commit/d237c3816588004628bd7d7… |
| 22:39:31 |
C0AT4G337UM |
:white_check_mark: *Readwise MCP* has moved from *Learn* → *Done*. Project complete!
This one ran from backlog (April 2025) through ship (April 2026) — Piotr built and expedited the full-featured MCP… |
| 20:26:07 |
C0ANR31CGQG |
:checkered_flag: *Bookstore Pages v2 → Done*
This project ran from October 2025 to June 2026 — from Dan's initial Book Pages IA framing through six design rounds, a Kris + Tadek build split, and a mo… |
| 20:24:50 |
C0B2RT1NYSY |
*Bookstore Sales Tax* is Done. Nice work <@U055PSQ2TCN> and <@U014GD6FE2C>!
From zero to production in 21 days — Anrok sales tax integration is live with agency-aware routing, per-jurisdiction PRH br… |
| 20:24:33 |
C0B1FQ9AZDW |
Moved *Bookwise Onboarding — Password Reset* to *Build*. Owner: <@U027D8YCD63>.
Shape was intentionally skipped — design pass deferred to Expedite.
*Scope entering Build:* Add a "Forgot password?" l… |
| 20:24:07 |
C0AUK4CPY2H |
:railway_car: *Bookwise Invite Codes → Expedite*
*What was built:* Mati shipped the full invite flow — referrer side in Settings, web landing page with dynamic OG images, invite-code-enforced signup.… |
| 20:23:40 |
C0B154X05EW |
Project moved to *Done*! Owner: <@U6NHLR96W>
*What was accomplished:* Fedor led three rounds of design work over three weeks, producing a comprehensive IA redesign — the "Add Button IA Master Project… |
| 04:00:59 |
C0B37G9KW9J |
:train: Update — Fedor delivered multi-book layout designs (up to 8 covers, finished books centered, overflow by words read); David sharing early build visuals. <https://github.com/readwiseio/rails/co… |
| 03:44:17 |
C0ASTN107FB |
:train: Update — emphasizeText correction step reviewed and approved, citation race condition fix shipped. Library search improvement is the last item before flag removal. <https://github.com/readwise… |
| 04:03:02 |
C0B0KE0P41L |
:train: Update — Artem opened PR #9639, rebuilding onboarding as an isolated navigation flow. Implementation of the designed self-serve flow is underway. <https://github.com/readwiseio/rails/commit/d6… |
| 04:01:33 |
C0APX8ZDR9S |
:train: Update — GIF persistence fixed (PRs #9628 + #9633 merged); seamless loading PR #9635 open; Kris delivered Expedite design doc, ruled out progressive blur/color extraction. Three of four origin… |
| 04:01:12 |
C0B154X05EW |
:train: Update — Shape complete! Fedor delivered his wrap-up (IA Master Project + Readerboard entry point), Dan confirmed design is done. Project ready for Build transition whenever you're ready, <@U6… |
| 04:01:09 |
C0AUK4CPY2H |
:train: Update — Invite codes now folded into broader onboarding; Artem integrating in parallel. Core build done, polish pass with Dan + Kris pending. <https://github.com/readwiseio/rails/commit/17b75… |
| 20:06:25 |
C0B1FQ9AZDW |
Hey <@U016ZPC564X> — this channel already has a project: *Bookwise Onboarding — Password Reset* (currently in Backlog, no owner). Happy to use that rather than creating a duplicate.
For the ticket: t… |
| 04:01:48 |
C0B0KE0P41L |
:train: Update — Design handoff meeting held. Figma specs added to project docs. Artem starting implementation of the self-serve onboarding flow. <https://github.com/readwiseio/rails/commit/6882b6c962… |
| 04:01:33 |
C0APX8ZDR9S |
:train: Update — David merged scroll-lock PR #9627; Kris planning progressive blur + adaptive defaults using existing color extraction. <https://github.com/readwiseio/rails/commit/6882b6c962dd15ef94c2… |
| 04:01:30 |
C0ASTN107FB |
:train: Update — emphasizeText regression found in global chat links, correction step proposed but streaming complicates implementation. Tristan open to scope revisit if Flash overloaded. <https://git… |
| 04:01:24 |
C0ASJFQ8G2Y |
:train: Update — Progressive loading shipped; latency drops from 500–1500ms to 100–500ms. Performance blocker cleared. Only minor UI polish remaining (title highlighting, content clipping). <https://g… |
| 04:01:07 |
C0AQWUHNMG8 |
:train: Update — Adam's Android BottomSheet fix (#9565) merged yesterday, clearing the last known blocker. <https://github.com/readwiseio/rails/commit/3b4aa7a98dcda8742f01f068a77ec73bfaf2cf90|view dig… |
| 04:02:13 |
C0B2FU8BDQC |
:train: Update — PR #9488 (bookstore KINDLE wire format) merged. The dedup/precedence stack is now end-to-end complete; minimize-bug and count-discrepancy remain. <https://github.com/readwiseio/rails/… |
| 04:01:59 |
C0B0KE0P41L |
:train: Update — Dan + Kris onboarding designs are complete. Handoff meeting with Artem today. Kindle wire format PR merged. Project shifting from pipeline work to implementation of the designed flow.… |
| 04:01:41 |
C0ASJFQ8G2Y |
:train: Update — Beta live on web + mobile for all users; prewarming shipped but perf still slow on broad queries (1500–2700ms). Progressive loading (top 20 first) proposed as next optimization. <http… |
| 04:01:31 |
C0ASTN107FB |
:train: Update — BYOK shipped to prod, alt models tested (no switch needed), citation race condition found and being fixed. Project near ship-ready. <https://github.com/readwiseio/rails/commit/0507f8c… |
| 04:01:24 |
C0APX8ZDR9S |
:train: Update — PR #9512 (image-aware foreground contrast) merged yesterday, closing the status-bar visibility bug. Remaining Expedite polish unchanged. <https://github.com/readwiseio/rails/commit/05… |
| 04:03:31 |
C0AQWUHNMG8 |
:train: Update — big catch-up since Apr 30: feature shipped to users, renamed to "Quick Lookup," tabs merged, Tristan on RAMP; early user feedback tracked in Notion, Android BottomSheet fix in flight.… |
| 04:01:21 |
C0ASJFQ8G2Y |
:train: Update — Web beta live for all users; skeleton loading state chosen; Christina's mobile design approved. Next: implement skeleton + ship mobile toggle. <https://github.com/readwiseio/rails/com… |
| 04:01:20 |
C0ASTN107FB |
:train: Update — reworked skills shipped to prod, BYOK PR #9575 opened. Two items remain (alt models + library search), neither blocking testing. <https://github.com/readwiseio/rails/commit/1441da2f90… |
| 04:00:46 |
C0B154X05EW |
:train: Update — Dan's feedback on the second design round is in via Notion. Shape converging — next up is Fedor incorporating feedback. <https://github.com/readwiseio/rails/commit/2e03402359bd4ff4f75… |
| 16:54:39 |
C0AQWUHNMG8 |
Got it — ownership transferred from <@U01JNR7BPV3> to <@U6MTAHZA4> for Ship. You're on point for RAMPing FCL out to users.
<https://github.com/readwiseio/rails/commit/f847588|view commit> |
| 16:53:15 |
C0AQWUHNMG8 |
:ship: *First Class Lookup → Ship*
Expedite is wrapped — moving FCL to Ship. Here's the handoff:
*What Expedite accomplished:*
• Adam resolved the bulk of the plate across two OTA rounds — auto-tab … |
| 04:01:21 |
C0ASJFQ8G2Y |
:train: Update — Christina designing "Better Search (beta)" toggle UX and loading state; rollout strategy settled on opt-in toggle, now live on prod search page. <https://github.com/readwiseio/rails/c… |
| 04:01:21 |
C0B154X05EW |
:train: Update — Fedor's second design round landed (Tabbar v2, Search v2, Move to shelf v2, Track, Weekly snapshot). Dan converting to product specs — Build transition likely once gaps are plugged. <… |
| 04:01:18 |
C0APX8ZDR9S |
:train: Update — PR #9512 still awaiting review (5 days); Dan confirmed default-bg needs full rethink for dark mode; flashing + drag nits flagged; Artem filed loading-state ticket. <https://github.com… |
| 04:01:10 |
C0ASTN107FB |
:train: Update — skills spec scope complete (PR #9465 updated, /trigger PR #9534 opened), back bug fixed, BYOK deferred. Tristan reviewing to unblock shipping. <https://github.com/readwiseio/rails/com… |
| 04:01:08 |
C0B2RT1NYSY |
:train: Update — FDC validated Anrok data flows. Remaining items are report polish: Macmillan format column + currency on free books, PRH official product ID, user full names. <https://github.com/read… |
| 19:19:01 |
C0B2RT1NYSY |
Moved *Bookstore Sales Tax* to *Expedite*. Owner: <@U055PSQ2TCN>.
Shape and Build collapsed into one sprint — Tadek's Anrok integration (PR #9346) went from proposal to production in 8 days. What's l… |
| 04:01:59 |
C0APX8ZDR9S |
:train: Update — David opened PR #9512 (image-aware foreground contrast via dominant color extraction); Dan flagged status-bar visibility bug that it likely fixes. <https://github.com/readwiseio/rails… |
| 04:01:47 |
C0B2FU8BDQC |
:train: Update — PR #9458 (cross-edition ASIN resolution) and PR #9482 (Kindle reset endpoint + Debug Settings) both merged. Dedup + precedence items solidly closed with recovery tooling in place; min… |
| 04:01:40 |
C0ASJFQ8G2Y |
:train: Update — Schema rolled to all 27k namespaces; v1 speed (~600ms) approved by Tristan; performance blocker cleared. Remaining: rollout strategy, ranking tuning, Bookwise ghost results. <https://… |
| 04:01:34 |
C0ASTN107FB |
:train: Update — skills PR #9465 opened for review, Dan testing (shortcut deferred post-flag-removal), wayfinding confirmed good by Kris. <https://github.com/readwiseio/rails/commit/bef9c9181677ed6912… |
| 04:01:30 |
C0B0KE0P41L |
:train: Update — Artem's ASIN-to-ebook resolution and Kindle hard-delete endpoint both merged. Kindle wire format PR in flight. Import pipeline steadily clearing. <https://github.com/readwiseio/rails/… |
| 18:10:51 |
C0ASJFQ8G2Y |
<https://github.com/readwiseio/rails/commit/b57b30802c886025d96f54718ca5f11c914f3a9d|view commit> |
| 18:10:50 |
C0ASJFQ8G2Y |
Moved *Reader Search via Turbopuffer* to *Expedite* :rocket:
*Build summary:* Mati shipped stemming/stopwords/ascii-folding, match highlighting UI, server search toggle, and `generated_summary` index… |
| 14:34:51 |
C0APX8ZDR9S |
Got it — I can see the backfill. Custom Profile Pages is now tracked in Expedite with <@U078C09MQQ5> as owner.
Here's where things stand:
• *Stage:* Expedite
• *Owner:* <@U078C09MQQ5> (transferred f… |
| 13:16:42 |
C0APX8ZDR9S |
Hey <@U6NHLR96W> — I don't have a project tracked for this channel yet. Could you tell me the project name so I can create it and move it straight into Expedite with <@U078C09MQQ5> as owner? Something… |
| 13:49:21 |
C0B1FQ9AZDW |
:train: Update — VIP author blocked at Bookwise signup by forgotten Readwise password; Dan flags password reset as near-future priority (still behind P0 onboarding work). <https://github.com/readwisei… |
| 09:46:43 |
C0ANR31CGQG |
Fixed — stripped out all the unrelated Bookwise infra activity (TrackedBook enforcement, cover thumbnails, canonical ebook resolution, group identifier, etc.) from the digests. Deleted the 05-18 and 0… |
| 09:44:06 |
C0ANR31CGQG |
Fair point — I was casting too wide a net. Cover thumbnails, canonical ebook resolution, TrackedBook enforcement — that's all general Bookwise infra, not this project. I'll keep the digest focused on … |
| 04:01:58 |
C0AT4G337UM |
:checkered_flag: *Readwise MCP* has moved from *Ship* → *Learn*.
*Owner:* <@U6MTAHZA4> (unchanged)
*Ship recap:* Coordinated launch with Piotr, Dan, and Christina — reframed the story around MCP as … |
| 04:01:50 |
C0B0KE0P41L |
:train: Update — Dan + Kris paired on self-serve onboarding design. Artem's GR/Kindle dedup matcher and photo-upload fix merged; ASIN-to-ebook resolution PR opened. Blockers flagged: Kindle import bro… |
| 04:01:38 |
C0ASTN107FB |
:train: Update — skills spec shared, sidebar animations + wayfinding shipped to prod/staging, retrieval fix split into prompt PR now + reranking Thursday. <https://github.com/readwiseio/rails/commit/5… |
| 04:01:37 |
C0ANR31CGQG |
:train: Update — cover thumbnails shipped end-to-end (Lior); Artem building canonical ebook resolution to fix wrong-cover issues. <https://github.com/readwiseio/rails/commit/501815b77d980f95d408b7b0f8… |
| 04:01:21 |
C0B2FU8BDQC |
:train: Update — PR #9398 (unified dedup + Kindle precedence) merged; Artem opened PR #9458 to resolve cross-edition ASINs to a canonical ebook. Two of four Expedite items done. <https://github.com/re… |
| 04:01:18 |
C0B2RT1NYSY |
:train: Update — PR #9346 merged and live. Anrok sales tax integration is in production — store credit taxed on gross retail price confirmed. FDC validating sales reports next. <https://github.com/rea… |
| 04:01:36 |
C0B2RT1NYSY |
:train: Update — FDC reviewed publisher report fields in PR #9346; Tadek pushed /bin/bash checkout support and sales-report end_date fix. A few report corrections remain before merge. <https://github.… |
| 04:01:05 |
C0ASTN107FB |
:train: Update — new retrieval bug found (domain/author search failing), Kris scoped PR #9415 to animations. <https://github.com/readwiseio/rails/commit/73fa434bdf33bcbfe0bffd0be27602af77d1591a|view d… |
| 04:02:09 |
C0B0KE0P41L |
:train: Update — Dan's full run-through feedback captured (May 15). Artem opened Find Friends suggestions (PR #9408) and GR/Kindle dedup (PR #9398); onboarding test tiles merged. Kris joining next wee… |
| 04:01:52 |
C0B2FU8BDQC |
:train: Update — Artem opened PR #9398: unified Kindle/Goodreads dedup matcher with Kindle precedence, addressing two of the four Expedite polish items. <https://github.com/readwiseio/rails/commit/e84… |
| 04:01:41 |
C0B2RT1NYSY |
:train: Update — Tadek opened full Anrok integration PR (rekindled #9346) covering agency-aware routing, PRH jurisdictional breakdowns, and refund handling. Awaiting review + finance input on store cr… |
| 04:01:33 |
C0ASTN107FB |
:train: Update — bugfixes shipped, wayfinding implementation underway (PR #9370), "Chat" vs "Ghostreader" naming debate between Tristan and Kris still open. <https://github.com/readwiseio/rails/commit… |
| 04:01:32 |
C0ANR31CGQG |
:train: Update — cover swap on reader open fixed (Artem); mobile cover resolution improved (Johannes). TrackedBook write-enforcement PR now two weeks open; cover color work not yet started in repo. <h… |
| 04:01:14 |
C0B154X05EW |
:train: Update — Fedor shared add button unbundling designs (Supercut + Figma prototype); Dan gave early feedback via Notion. Feedback loop active heading into next week. <https://github.com/readwisei… |
| 04:01:10 |
C0AUK4CPY2H |
:train: Update — Dan tested and approved the core build; Mati shipped dynamic OG images. Ball now with Dan + Kris to polish the experience. <https://github.com/readwiseio/rails/commit/c456279|view dig… |
| 04:01:40 |
C0B0KE0P41L |
:train: Update — Artem's onboarding-cards fix (PR #9313) merged; import-source spacing equalized (PR #9339). UI reorder is the last major expedite item. <https://github.com/readwiseio/rails/commit/50d… |
| 04:01:21 |
C0AUK4CPY2H |
:train: Update — Mati shipped the referrer flow (staff-only) + web landing page, now building referee onboarding. Play Store public move in scope post-ship. <https://github.com/readwiseio/rails/commit… |
| 04:00:57 |
C0B37G9KW9J |
:train: Update — Dan clarified monthly finished-books layout (full-size first, shrink to fit); no PRs yet, Build ramping up. <https://github.com/readwiseio/rails/commit/b7b783216708540f6c62fbba02e0283… |
| 00:20:03 |
C0B0KE0P41L |
You're right, sorry about that. Ibai's and Rasul's PRs are separate workstreams — I was matching on keywords like "onboarding" and "Kindle" too loosely. I've scrubbed all their wrongly attributed entr… |
| 00:00:15 |
C0ASTN107FB |
:train: Update — Piotr fixed most bugs and PR #9265 ready to ship, Christina delivered /skillname and settings designs, pace picking up toward Friday target. <https://github.com/readwiseio/rails/commi… |
| 04:04:27 |
C0AUK4CPY2H |
:railway_car: *Bookwise Invite Codes → Build*
Shape is wrapped. Here's the handoff:
*What was shaped:* Single invite code per user, shareable, claimable up to 5 times. Positioning is "discover your … |
| 04:01:58 |
C0B2RT1NYSY |
:train: Update — Tadek recommends direct Anrok API integration (Option A); shipping confirmed tax-free; three store-credit edge cases awaiting finance input. <https://github.com/readwiseio/rails/commi… |
| 04:01:57 |
C0B0KE0P41L |
:train: Update — Ibai polished onboarding + import screens; Artem fixing onboarding-cards disappearance bug from step removal; Rasul's Kindle mixed-language import fix merged. Import Sources UI reorde… |
| 04:01:47 |
C0ASTN107FB |
:train: Update — Kris delivered wayfinding design pass (no peek, wider sidebar, tab rename to Ghostreader), Christina started settings design, Piotr deployed warm/cold + fixes to staging. <https://git… |
| 04:01:35 |
C0ANR31CGQG |
:train: Update — Hannes's group identifier fix merged (improves cover selection); Artem fixed wrong-book iOS bug; TrackedBook write-enforcement PR still open. <https://github.com/readwiseio/rails/comm… |
| 20:07:23 |
C0B37G9KW9J |
Project created: *Bookwise Snapshot Stickers* (<https://github.com/readwiseio/rails/commit/427e702|view commit>)
• *Stage:* Build
• *Owner:* <@U0B19NJ29E1>
• *Initiative:* bookwise-launch
I've captu… |
| 20:03:44 |
C0B37G9KW9J |
Happy to create a project! A couple quick questions:
1. What's the project name?
2. One sentence on what it's about (this shows up on the dashboard card)
3. Who's the owner? |
| 19:44:21 |
C0B2RT1NYSY |
Project created! :white_check_mark:
*Bookstore Sales Tax* is live in Shape.
- *Owner:* <@U055PSQ2TCN>
- *Initiative:* Bookwise Launch
- *Stage:* Shape
<@U055PSQ2TCN> — you're up. The thread above h… |
| 19:43:13 |
C0B2RT1NYSY |
Hey <@U6MTAHZA4> — I see FDC's summary on the Stripe/Anrok tax product ID setup. This channel isn't linked to a project yet though, so I don't have much context beyond what's in this thread.
What do … |
| 04:01:44 |
C0ASTN107FB |
:train: Update — GG deprioritized behind bookstore context transfer; RFD review from Kris still pending, Tristan flagged as potential blocker for Piotr. <https://github.com/readwiseio/rails/commit/0a3… |
| 04:05:51 |
C0ASJFQ8G2Y |
:train: Update — Quality assessed as nearly prod-ready; speed is the last blocker. Next step: engage the Turbopuffer team for performance optimization. <https://github.com/readwiseio/rails/commit/75a2… |
| 20:25:15 |
C0B2FU8BDQC |
My bad — you're right. Ibai's and Rasul's Kindle PRs (#9258, #9209) are for the original Readwise product, not Bookwise. I've removed them from this project's digest and documents. Will be more carefu… |
| 04:19:14 |
C0ANR31CGQG |
:train: Update — sticky nav z-index fix merged; Hannes fixing group identifier for better cover selection; FDC shared legal docs draft for bookstore integration. <https://github.com/readwiseio/rails/c… |
| 04:19:06 |
C0B0KE0P41L |
:train: Update — PR #9061 (GR client-side scrape) merged, clearing the critical path for Import Sources reorder. Ibai also landed Kindle sync fix + configurable import sources. <https://github.com/rea… |
| 04:18:48 |
C0B2FU8BDQC |
:train: Update — Ibai's Kindle sync fix + import-source config merged (PR #9258); Rasul's roman-pages fix still in review. <https://github.com/readwiseio/rails/commit/bb50dd0aa850451d8c9cb76cc61617c80… |
| 04:18:23 |
C0ASJFQ8G2Y |
:train: Update — Mati shipped URL/description match highlighting in search results UI; Bookwise ghost results identified as a new edge case to handle. <https://github.com/readwiseio/rails/commit/94078… |
| 22:30:38 |
C0B1FQ9AZDW |
Nope, nothing to file — Dan was just getting me linked to this channel. Thanks for checking! |
| 21:02:44 |
C0B1FQ9AZDW |
My bad — the channel wasn't linked in the project config. Fixed now, I'm wired up to *bwl-new-onboarding-password-reset* (<https://github.com/readwiseio/rails/commit/bde6b4e|view commit>).
Project's … |
| 20:54:52 |
C0B1FQ9AZDW |
Yep, I'm here! Which backfill are you referring to — a specific project, or something new you'd like me to set up? |
| 18:19:56 |
C0B154X05EW |
Done — linked to the Bookwise Launch initiative. (<https://github.com/readwiseio/rails/commit/0a43413|view commit>) |
| 17:58:00 |
C0B2ET7MCP5 |
Project created and tracking! Here's the setup:
*Bookwise Creator Brief*
• *Stage:* Shape
• *Owner:* <@U021M0A9CSW>
• *Initiative:* Bookwise Launch
• *Description:* Preparing a brief for ~100 authent… |
| 17:54:41 |
C0B2ET7MCP5 |
Elasticsearch is down so I can't read the channel history right now. No worries — just give me the basics and I'll get this set up:
1. *Project name* — what should I call it?
2. *One-sentence descrip… |
| 17:50:58 |
C0B154X05EW |
Updated — channel name set to #bwl-add-button-unfucking. (<https://github.com/readwiseio/rails/commit/40173c6|view commit>) |
| 17:44:21 |
C0B2ET7MCP5 |
Hey <@U6NHLR96W>! Happy to start tracking this project. A couple of quick questions so I can set it up:
1. What's the project name?
2. Can you give me a one-sentence description of what it's about?
3… |
| 17:35:33 |
C0B154X05EW |
Project created and tracking! Here's what I set up:
*Add Button Unfucking* — now in *Shape*
Owner: <@U6NHLR96W>
I captured the Loom brief as the first Shape document and wrote up the motivation from… |
| 17:32:15 |
C0B154X05EW |
Hey <@U6NHLR96W>! Happy to start tracking this project. I need a couple of details:
1. *What's the project name?* (I'll derive the slug from it)
2. *A one-sentence description* — this shows up on the… |
| 04:28:53 |
C0ASJFQ8G2Y |
Good catch — removed the Ghostreader eval references from this project's digest. That belongs to the global-ghostreader project. Updated digest now focuses on the team's search testing feedback only. … |
| 04:04:17 |
C0B0KE0P41L |
:train: Update — Ibai's welcome-screen animations merged, Artem landed GR hard-delete admin tool; GR client-side scrape (PR #9061) is the critical-path item still in flight. <https://github.com/readwi… |
| 04:01:58 |
C0ANR31CGQG |
:train: Update — fast-fix batch shipped, community highlights cache merged (perf fix), cover selection algo flagged for improvement; Kris moving to cover colors next. <https://github.com/readwiseio/ra… |
| 04:01:13 |
C0ASJFQ8G2Y |
:train: Update — Stemming/stopwords/ascii-folding deployed to test profiles; Ghostreader evals improved 70→77%. Dan and Tristan confirm search is better than SQLite; ranking and performance are the ne… |
| 04:00:45 |
C0B2FU8BDQC |
Yep, I'm here! Backfill landed — I've got the full history from Dan's original Fable holy-shit moment through Shape, Build, and into Expedite. Project is tracked as *Bookwise Onboarding — Kindle Impor… |
| 04:00:40 |
C0B0KE0P41L |
Yep, I'm here! Backfill is in — project is tracked at expedite with <@U016ZPC564X> owning. Yesterday's digest captured your BIG PICTURE brief (drop first reading session, Goodreads primary CTA, Kindle… |
| 23:11:35 |
C0AQCHP3N7M |
I can't create Linear tickets right now — the contextone MCP needs to be authenticated first and I don't have a Linear API key configured. Someone would need to either:
1. Approve the contextone OAut… |
| 23:10:22 |
C0AQCHP3N7M |
Sure — here's what I'm thinking based on the thread:
*What Artem wants:* A way to derive Linear projects and tickets from Rails project data (Slack channels, stage transitions, etc.), with custom ann… |
| 20:58:11 |
C0AQCHP3N7M |
Honestly, no — I can't resolve channel IDs to names on my own. I see messages tagged with `C0AQCHP3N7M` but I don't know the human-readable channel name. What channel is this? |
| 20:57:29 |
C0AQH5QB9NZ |
I don't see a project for "new onboarding registration" in git. The current projects are:
bwl-bookstore-pages-v2, bwl-feed-post-for-updates, bwl-invite-codes, first-class-lookup, global-ghostreader, … |
| 20:57:07 |
C0AQCHP3N7M |
Ha, fair point — I shouldn't be guessing, you should be telling me! 😄
This channel isn't mapped to any project right now. What's it for? I can either:
• Create a new project for it
• Point an existi… |
| 20:56:44 |
C0AQH5QB9NZ |
Yep, I'm here! 👋 What backfill are you referring to? I don't see this channel linked to a specific project — is there something you'd like me to set up or look into? |
| 20:56:27 |
C0AQCHP3N7M |
That's because there's no project linked to this channel yet — none of the project `state.yml` files have this channel ID. Which project should be associated with this channel? I can either create a n… |
| 20:50:33 |
C0AQCHP3N7M |
Yep, I'm here! I can see the backfill — `bwl-bookstore-pages-v2` was backfilled at expedite stage, and there have been a few auto-updates and digest refreshes since then. Also got `reader-search-via-t… |
| 04:02:18 |
C0ANR31CGQG |
:train: Update — Artem's TrackedBook refactor is underway on rekindled (PR #9151 merged + write-enforcement branch active); Kris's fast-fix batch (variant swap, publisher/ISBN, sticky cover) posted fo… |
| 04:02:12 |
C0ASTN107FB |
:train: Update — RFD shared with Kris, who flagged global-to-doc transition and skills/legacy prompt integration as key design areas. Piotr addressing warm/cold feedback and starting bugfixes. <https:… |
| 04:01:23 |
C0ASJFQ8G2Y |
:train: Update — Now in Build; stemming fix confirmed trivial (schema update auto-re-tokenizes, no re-index/version bump needed). Next: deploy stemming change and test cross-product impact. <https://g… |
| 21:37:26 |
C0ANR31CGQG |
Done — pulled your latest changes. The dashboard should pick them up now. |
| 19:56:37 |
C0ASJFQ8G2Y |
<https://github.com/readwiseio/rails/commit/4c65d13cdee213b867dec9ce5a0af5368ca1a65b|view commit> |
| 19:56:29 |
C0ASJFQ8G2Y |
Moved *Reader Search via Turbopuffer* to *Build* :hammer:
*Shape summary:* Mati proved the plumbing works — POC with v1/v2 server variants, dev branch toggle, and mobile build 8.99.1. Dan and Tristan… |
| 19:46:06 |
C0ANR31CGQG |
Pulled the latest — found it! I'm now connected to *Bookstore Pages v2*.
Here's where things stand:
• *Stage:* Expedite
• *Owner:* <@U078C09MQQ5>
• Post-launch polish underway on `bookstore-v2-fast-f… |
| 19:45:28 |
C0ANR31CGQG |
Hey <@U6NHLR96W> <@U6MTAHZA4> — I'm here and listening! 👋
I don't have a project connected to this channel yet. A couple of options:
1. *Create a new project* — tell me the project name and a one-s… |
| 04:01:19 |
C0ASTN107FB |
:train: Update — Tristan expanding punch list with product improvements, drafting final RFD, seeking design support from Kris/Christina this week. <https://github.com/readwiseio/rails/commit/1caad88f6… |
| 04:01:17 |
C0AUK4CPY2H |
:train: Update — Jackson posted round 2 of the referrer flow (Loom + Notion) with tweaks from team feedback; added docs to the project. <https://github.com/readwiseio/rails/commit/c8af029|view digest> |
| 04:01:05 |
C0ASJFQ8G2Y |
:train: Update — Dan and Tristan both flagged search quality gaps in the POC (phrase matching, content-type queries); core debate shifting from architecture to result quality. <https://github.com/read… |
| 15:48:08 |
C0ASTN107FB |
Global Ghostreader moved to *Expedite*. Owner: <@U6MTAHZA4>.
*What Build delivered:* Full agentic chat (M1–M6), skills in prod, eval harness, system prompt iteration, warm/cold orbit system (in revie… |
| 04:01:52 |
C0ASTN107FB |
:train: Update — system prompt PR shipped, eval tpuf bug root-caused, Adam reviewed warm/cold, Tristan's bug report filed with Linear tickets. <https://github.com/readwiseio/rails/commit/07b0e83dfe179… |
| 04:01:28 |
C0AQWUHNMG8 |
:train: Update — error-state button behavior settled; Dan flagged dictionary speed and verbose non-standard term responses as final expedite items. <https://github.com/readwiseio/rails/commit/03c7cfc|… |
| 04:01:20 |
C0AUK4CPY2H |
:train: Update — Kris reviewed Jackson's latest Figma round: referrer mechanics tracking, "why invite" framing is the open design question, and address book trending toward "later." <https://github.co… |
| 04:01:12 |
C0ASJFQ8G2Y |
:train: Update — Mati pushed a working Turbopuffer search POC with v1/v2 A/B testing to his dev branch and mobile build 8.99.1; content-type matching still WIP. <https://github.com/readwiseio/rails/co… |
| 00:43:43 |
C0ASTN107FB |
Done — <@U07TFNXANMB> is the owner again. <https://github.com/readwiseio/rails/commit/5cada0077b757a01878c593b4fbe70d1c90d92e3|view commit> |
| 04:02:02 |
C0ASTN107FB |
Done — you're the owner now. <@U07TFNXANMB> stays on for the warm/cold PR (#8861) with Adam's review, but you're driving from here.
Updated state.yml, digest, and stage summary. <https://github.com/r… |
| 04:01:56 |
C0ASTN107FB |
:train: Update — ownership transferred to <@U6MTAHZA4>; first hands-on testing very positive, planning expedite and legacy skills migration. <https://github.com/readwiseio/rails/commit/51ecc6c44419b58… |
| 21:43:05 |
C0AUK4CPY2H |
All changes pushed — <https://github.com/readwiseio/rails/commit/f51dbe3|view commit> |
| 21:41:07 |
C0AUK4CPY2H |
Done — backlog start date set to April 21 and summary updated with your text. |
| 21:37:31 |
C0AUK4CPY2H |
Fixed. The backlog summary was carrying fabricated details from the original migration — stripped all of that. It now just says what actually happened: queued up as the next growth project, moved quic… |
| 21:34:35 |
C0AUK4CPY2H |
Done — two changes committed (619eef3, push pending auth fix):
*Stage summaries* — rewritten to ground in your Cold Start Problem framing instead of the earlier guessed context. Backlog summary now l… |
| 21:30:59 |
C0AQWUHNMG8 |
Yep, I'm here! What's up? |
| 21:30:06 |
C0APWQA9E1E |
Good catch — added "session health alerting" to the Expedite to-do list. That's exactly the kind of thing that needs solving before wider rollout.
Ironically, my own git auth just expired so I can't … |
| 21:29:57 |
C0AUK4CPY2H |
Done — motivation.md is updated with your version. Commit is local (442d867) but push is failing due to a GitHub auth issue; it'll go out once the token is refreshed.
Good idea workshopping these doc… |
| 05:00:10 |
C0ASTNPBGTH |
Hey <@U6MTAHZA4> — I'm on it! I'm tracking *RevenueCat Migration* now.
Here's where things stand:
• *Stage:* Build · *Owner:* you
• *Initiative:* finish-readwise-wip
• Reader Android and Reader iOS … |
| 04:56:04 |
C0ASJFQ8G2Y |
Already tracking *Reader Search via Turbopuffer* — here's where things stand:
• *Stage:* Shape
• *Owner:* <@U027D8YCD63>
• *Initiative:* finish-readwise-wip
Mati is doing technical exploration and P… |
| 04:01:17 |
C0ASTN107FB |
:train: Update — Tristan opened first eval case (Evan Spiegel link-preservation regression, PR #9060). Eval system confirmed to work for doc-level chats too. <https://github.com/readwiseio/rails/commi… |
| 04:01:13 |
C0AQWUHNMG8 |
:train: Update — Dan and Tristan report FCL looking sharp after Adam's latest OTA; model speed and offline fallback now on the table. <https://github.com/readwiseio/rails/commit/6fa727d|view digest> |
| 04:01:36 |
C0AUK4CPY2H |
:train: Update — Tristan left Figma feedback; address-book research came back inconclusive, leaving scope call with Dan. <https://github.com/readwiseio/rails/commit/36e4646|view digest> |
| 21:45:32 |
C0AQWUHNMG8 |
:train: Update — DRM testing surfaced X-Ray breakage + triggered rearchitect of Define/Lookup context sourcing; Adam OTA'd bulk fixes and auto-tab classification. <https://github.com/readwiseio/rails/… |
| 21:45:13 |
C0ASTN107FB |
:train: Update — Skills PR #8863 merged to prod. Eval skill diagnosed title-search retrieval gap from Tristan's penicillin test and auto-created fix PR #9022. Evals PR #8898 up for review. <https://gi… |
| 21:44:53 |
C0AUK4CPY2H |
Fair point — you're right. <@U6NHLR96W> owns Shape, so the next step is his call: land the scope (address book in or out), incorporate Figma feedback, and write up the shape doc. Your address-book res… |
| 21:44:19 |
C0AUK4CPY2H |
<@U6NHLR96W> owns Shape. Next step: Tristan's address-book research — once you have a read on the current "meta" for address-book patterns, that'll close the last open question and we can write up the… |
| 21:42:37 |
C0AUK4CPY2H |
I'm here! 👋 Caught up on Bookwise Invite Codes — we're in *Shape* with <@U6NHLR96W> owning. What do you need? |
| 10:44:18 |
C0AQWUHNMG8 |
Updated the expedite doc with to-dos from the latest conversation (<https://github.com/readwiseio/rails/commit/2ceac9f|view commit>). Here's what I added:
• *Disable X-Ray tab for DRM books* — keep t… |
| 12:51:14 |
C0APWQA9E1E |
Added to open questions in expedite.md — responsive mobile design, with your leaning toward doing it later. <https://github.com/readwiseio/rails/commit/5b00f36|view commit> |
| 04:01:31 |
C0APWQA9E1E |
:train: Update — Dan's Expedite to-dos complete (PRs #6-8), ownership back to Tristan, avatar set. Two items left: dashboard inspection and channel deployment. <https://github.com/readwiseio/rails/com… |
| 01:07:33 |
C0APWQA9E1E |
Looking good! Checked off the avatar. Two to-dos left. <https://github.com/readwiseio/rails/commit/349c43f|view commit> |
| 01:04:12 |
C0APWQA9E1E |
Done — ownership transferred from <@U6NHLR96W> to <@U6MTAHZA4>. Current stage emphasis checked off.
Remaining to-dos are on <@U6MTAHZA4>:
• Inspect active projects on the dashboard, flag gaps
• Deplo… |
| 01:01:45 |
C0APWQA9E1E |
Francis Bourgeois is perfect. A trainspotter who gets genuinely ecstatic about trains as the face of a project management bot called RAILS — that's hard to beat. The GoPro facecam angle is instantly r… |
| 00:50:34 |
C0APWQA9E1E |
Some ideas riffing on the train/rails theme + playful company vibe:
• *Conductor hat on a clipboard* — project management meets trains, slightly absurd
• *A train emoji face* — like the 🚂 but with e… |