Maybe you're experiencing normal usage rates now that the 2x March promotion is over?
> From March 13, 2026 through March 28, 2026, your five-hour usage is doubled during off-peak hours (outside 8 AM-2 PM ET / 5-11 AM PT / 12-6 PM GMT) on weekdays). Usage remains unchanged from 8 AM-2 PM ET / 5-11 AM PT / 12-6 PM GMT on weekdays.
Source: https://support.claude.com/en/articles/14063676-claude-march...
Sadly this is still the case. Cloud Run helped us get off the ground. But we've had two outages where Google Enhanced Support could give us no suggestion other than "increase the maximum instances" (not minimum instances). We were doing something like 13 requests/min on this instance at the time. The resource utilization looked just fine. But somehow we had a blip in any containers being available. It even dropped below our min containers. The fix was to manually redeploy the latest revision.
We're now investigating moving to Kubernetes where we will have more control over our destiny. Thankfully a couple people on the team have experience with this.
Something like this never happened with Fargate in the years my previous team had used that.
Very cool! I built a no-code version of this for my friends' birthdays where all of the clues were about them (inside jokes, dumb bets, catch phrases, etc). I just printed out categories in a grid on a sheet of paper and had a bunch of questions listed out in OneNote on my phone and manually kept score. I didn't put penalties in place if somebody got it wrong to encourage people to participate vs just optimizing for score. "Jonpardy" and "Joshpardy" are two of our friend groups most cherished and repeatedly cited events from the past 10 years.
I've dabbled with the idea of creating some real implementation of this. I most recently put together an @Midnight / After Midnight no-code game for a friend's bachelor party. The buzz-in mechanics and score-keeping are somewhat similar. Maybe you could repurpose your core logic for several different game show themes?
It's pretty rare to give someone 10%+ equity after a business has been operating for several years. Even if you brought on a professional CEO, I wouldn't expect to see that. I've seen some founders put together special agreements to transfer some of their own shares to a person recognizing that they'll be taking over a big part of their own role and that it wouldn't be fair to dilute the rest of the company.
Regardless of what you do, make sure there is a vesting schedule in place (e.g. a 4 year vest + 1 year cliff).
For anybody else who also hadn't come across the term RVC before:
"The RVC model is a Retrieval-based Voice Conversion system using AI for high-quality voice cloning. It utilizes artificial intelligence to modify or clone voices in real-time."
Source: https://speechify.com/blog/rvc-vocal-models/
I enjoyed this blog. I think it provides a great succinct overview of various approaches native to Postgres.
For the "capture changes in an audit table" section, I've had good experiences at a previous company with the Temporal Tables pattern. Unlike other major RDBMS vendors, it's not built into Postgres itself, but there's a simple pattern [1] you can leverage with a SQL function.
This allows you to see a table's state as of a specific point in time. Some sample use cases:
- "What was this user's configuration on Aug 12?"
- "How many records were unprocessed at 11:55pm last night?"
- "Show me the diff on feature flags between now and a week ago"
spoiler: Of the 6 cars they tested, Tesla Model Y had the best performance in terms of miles per kWh and total range. But it still clocked in at only 81% of claimed range.
I rarely have to use Superchargers because I have charging at home. But when I do, it's about $25 for an 80% top up (e.g. going from 10% -> 90%) in my Model Y.
Just looked it up: $0.45/kWh the last 4 times at Tesla Superchargers in California.
I don't use Svelte, but I appreciate the principled approach and the clear explanation.
I primarily coded in Python for 4 years. Then I founded a company that didn't need as much data science and my primary language switched to JavaScript (2 years) then TypeScript (4 years). Overall, I really like TypeScript. But I do absolutely miss being able to cmd + click into a function/class from an open source package and see the actual implementation and not just a type definition. This is probably the biggest day-to-day frustration I have with TS compared to JS and Python.
I was a huge fan of Scrabble growing up. I like how simple this is and would probably be fun playing with my mom daily for a bit (similar to Wordle).
I like your game for what it is, but anchoring it to Scrabble confused me a bit when I first played. It was missing a couple of keys aspects for me:
1) Bingo bonus (+50 for using all of your tiles)
2) Optimizing your tileset so that future turns have a better chance for a bingo.
Also, minor nit: the colors for the letter multipliers (purple and red) in your game also confused me because in Scrabble the red x3 is a word multiplier and not a letter multiplier. You could consider using light blue and dark blue if you want to align to the traditional game.
I created a callbox bot using Twilio about 10 years ago. It started as something simple: when my callbox bot is called, play dial tone 9 (to signal the callbox to open the door) and then send me a text message that someone is coming up. I wasn't too worried about people abusing the system and this worked really well.
Over the years, I've made it more complex. I added a simple menu system using text-to-speech and accepting the DTMF tones for the selection. This way I could tell people to enter a secret PIN and they could provision themselves access to my lobby. If they didn't know the secret PIN, there was an option to have it patch through to my phone so I could screen the guest using normal voice and I could manually press 9 on my cell phone.
It also has an SMS interface where I can select different modes and hours of operations. So when I host a poker night, I'd txt my bot "poker" and it would it update the menu system to be Poker themed so that when guests were at the callbox they'd be greeted with poker puns. When my friend comes to visit from Austin, I'd txt "StarCraft" and he'd get to hear some dumb SC2 puns. I have 20+ different modes now for various occasions over the years. Now when people visit they get their phones out and record the greeting half-expecting something that they would want to remember or share on social media.
Sadly, my building just replaced their callbox with a newer model that does not accept secondary input. So once the call is connected to Twilio, if the user touches any button on the callbox (e.g. to enter a secret PIN), it will disconnect the call. I suppose it's now time to use some voice-to-text options to bring back the interactivity, but I suspect the lag would make the experience more frustrating than fun.
I've played with several of these Stable Diffusion frameworks and followed many tutorials and imaginAIry fit my workflow the best. I actually wrote Bryce a thank you email in December after I made an advent calendar for my wife. Super excited to see continued development here to make this approachable to people who are familiar with Python, but don't want to deal with a lot of the overhead of building and configuring SD pipelines.
There's a 2 week Whisper Fine Tuning event put on by Hugging Faces going on right now [1]. The new "large-v2" model was announced as part of the kickoff [2]. It's supposed to offer 5% - 10% improvements over the previous large model.
This reminded me of a project from long ago. When I was growing up, I was not particularly good about expressing gratitude to my family. I recognized this in myself and in college and wanted to make up for lost time.
I went around and interviewed my extended family and some of my mom's friends over the course of about 6 months. I asked each person for any fun or touching stories they had. There were a lot. I edited down many hours of footage into a 45 minute video that basically said, "thanks for the last 22 years".
Days before I moved across country to start my job after graduation, I got the family together and had a viewing party. Unbeknownst to my mom this wasn't a going away party for me, it was an appreciation party for her. It was a wonderful moment.
Two people who I interviewed as part of this project have since passed away. So I dug up and digitized all of the unedited footage so that I can easily pull it up on my computer. There are some wonderful memories from my grandma and my father at a happier/healthier time in their lives. In their final years they were not really able to speak at all, so this footage really captures them at their relative best. It has been a helpful way for me and my family to deal with grief.
OP: Thank you for sharing your story. Sorry for your loss.
> From March 13, 2026 through March 28, 2026, your five-hour usage is doubled during off-peak hours (outside 8 AM-2 PM ET / 5-11 AM PT / 12-6 PM GMT) on weekdays). Usage remains unchanged from 8 AM-2 PM ET / 5-11 AM PT / 12-6 PM GMT on weekdays. Source: https://support.claude.com/en/articles/14063676-claude-march...