Cole Palmer

Here is the new AWS State Machine that gathers all of the data for HockeyStatCards. 1. In the morning, a scheduled job runs (AWS Event Bridge) that gets all the games for that day, and schedules a poll start time 2ish hours after puck drop 2. We check if the status is final, if not, we have dynamic backoffs based on how much time is left in the game. We loop until finished. If we hit our max poll attempts, we error here. 3. If we get a final status, we run the compute lambda. 4. We then check a boolean if we should post on socials or not, if true, we post. Visually it looks like this:

question from our user research department

todays adventures - move all @profbstatcards and @cfbstatcards client side percentile calculations into the data processing layer

part of me wonders if local development as a whole will go away everything in the cloud ssh localhost, and scripts all in the cloud always on. close your laptop or phone anytime.

With an especially crowded Norris field this season, it's high time that the NHL creates a defenseman award that specifically rewards defensive excellence. Here's who would be in the running for that long overdue award this season. nytimes.com/athletic/71472…

@hockeystatcards ducks legend Alex Killorn

Current trending players on HockeyStatCards

hey hey not bad perf

@beicksquad79219 what do you mean

@cepvi0 hey man, can i not search by teams anymore for hockey? Desperately trying to figure out how!

@LLMJunky @theo is this merged in the original or in a fork?

God I love Open Source. Integrated browser into T3Code Gonna add a terminal next.

shipped a fix for our trending players bug today. taced to noisy view tracking and an overly complex cache/query path... so I simplified the whole flow: - trending now uses upstash hourly buckets - we record exactly one view when a player page mounts, and we compute “trending” by summing the last 24 hourly buckets - ttl-based cleanup so old keys expire automatically. should be more reliable trending data. always better with less moving parts

@APrivitelli conditional skills?

@cepvi0 I’m now using docs.conductor.build

Whats the best workflow for this: Im using the `superpowers` skills in codex, which is great for large features or big changes. but when i say adjust padding on X, I really dont need brainstorming, 3 different design options. So then I find myself going to Cursor for quick things and Codex for large features or bugs. Whats the best way to manage this all in one place? (conditional skills?)

@LuigiLemieux33 lol all good. just feel like this is such a layup problem to solve

@cepvi0 This is true. I had “doesn’t answer your question” at the start but was getting cut off haha. I just use ai a ton for coding and love it and felt compelled to give my opinion

@LuigiLemieux33 still doesn’t solve conditional skills in one harness though

@cepvi0 Claude code tends to stick to skills better I fine but Codex for large changes and Claude code using /chrome for front end fixes or visual things. Claude code for big changes is seemingly much worse than codex. Codex find so many bugs in Claude codes work. I have found.

@Leevon_L @hockeystatcards oh ok, thought you had the magic app

@cepvi0 @hockeystatcards Just coding with different skills. Seeing how it performs

who codes with ai tools? need help 👇

@Leevon_L @hockeystatcards trying out what?

@hockeystatcards I’ve been trying this out

@PastTheEyeTest good to hear. i think someone should build the functionality to optionally toggle off a skill or groups of skills. @theo thoughts on this for T3 code maybe?

@cepvi0 I don't think there's one AI tool other than maybe Claude Code that can handle all changes, I think the split workflow that you have b/w codex and cursor works I mainly use claude code though but from my experience there's not a single tool that does both automatically

@benchrates 🙌

@cepvi0 Yes, we will do it in one of the next releases, thanks for reminding me, Cole!

TOP NHL defensemen by xGF% after Olympic break | 5v5 without EN 1. Mattias Ekholm #LetsGoOilers 65.96 2. Cale Makar #GoAvsGo 64.92 3. Brayden McNabb #VegasBorn 63.28 4. Evan Bouchard #LetsGoOilers 63.18 5. Lian Bichsel #TexasHockey 63.10 6. Tyler Tucker #StlBlues 62.84 7. Lane Hutson #GoHabsGo 62.56 8. Thomas Harley #TexasHockey 62.41 9. Colton Parayko #StlBlues 61.66 10. Jordan Spence #GoSensGo 61.62 min. 90 minutes #NHL #hockeytwt

@TopDownHockey @JFreshHockey about time

#FlyTogether are basically playoff locks after beating #TuskUp last night. 🦆🔐 hockeystats.com/playoff-odds

@tyler__palmer incoming polymarket for which model is right for X

Today I asked Gemini, Grok, and Chat GPT to do deep research and estimate the day Trump declares victory and ends the War in Iran. Gemini - April 1st Grok - April 2nd Chat GPT - April 7th Interested to see who is closest.