Claude Code leaked their source map, effectively giving you a look into the codebase.
I immediately went for the one thing that mattered: spinner verbs
There are 187
We offered 5 people a Porsche 911 GT3 RS if they could get @WisprFlow to make a mistake
It's the fastest and most accurate AI voice dictation app that's 3x more accurate than ChatGPT, Claude, or Siri.
Today, we’re finally launching on Android. Download now: https://t.co/TJhnUhDSLv
As a part of the launch, we’re giving away 6 months of Wispr Flow Pro for free.
Like, retweet and comment ‘Wispr Flow’ to get it. Enjoy.
— Written with Wispr Flow
@SumitM_X A real world example of this I discovered is the funnel used to control water flow. The top opening limits the maximum capacity of the funnel, spilling over any extra water, while the bottom opening controls the speed at the water is released.
don't ask interviewers about "work-life balance", it's a very lazy question and no one is going to tell you it sucks.
ask about the on-call rotation. ask about the last page. ask how long it takes from an incident until the PIR. ask if the team prioritizes reducing pager disruptions. ask how quickly you need to respond to weekend emails.
that will give you most of the answers you're looking for.
@asmah2107 Stop read/write operations from monolith db gradually. Once all domains are fully migrated, retire the monolith database.
This allows easy rollback and good testing of individual microservices before replacing the monolith entirely. 2/2
@asmah2107 We could start by splitting the monolith into services step by step that initially share the same monolith database helping us validate each service’s domain logic independently, then use CDC build read models. In the next phase introduce a dual-write mechanism for both dbs. 1/2
@asmah2107 We could extract the IP and put it inside the RedisBloom using the minute level bucketing. Set a TTL of 1 hour so the in memory store isn't overwhelmed. Meanwhile a background job merges the last 60 filters, estimates cardinality, and returns unique visitor count.
@asmah2107 Google/Okta logins prevent this by smart use of 2FA triggers. Leverage fingerprinting/behaviour to detect unusual sign in. Could also layer this with CAPTCHA. A good case study for this would be to understand how Okta logins work.
https://t.co/H3rtXwVslk
@asmah2107 This reminds me of a classic thundering herd or celebrity symbol problem,where a couple of high traffic stocks overwhelm the system.Give top symbols their own shards; bundle the rest.
This allows better scaling & throughput.
It's finally here! Untitled UI React is live ✨
Untitled UI React is the world’s largest collection of open-source React components built with Tailwind CSS v4.1 and React Aria.
Just copy, paste, and build. 👇
@asmah2107 Whe using Postgres - user writes propagate to primary and replicas via WAL and LSN checkpoints. When a region goes down failover triggers and LSNs across regions are compared to detect inconsistencies. Recovery could be done using WAL replay.
@mozedev@asmah2107 That might work too but chances of reading the same data is very high after write so this might make the initial read a bit slow, otherwise this might work too
@agustingomes@asmah2107 A simple solution to this could be adding a retries for cache udpate failures. For a more scalable approach, we could introduce a queue to buffer the write requests and wait for Redis update to ack before dequeuing.
@agustingomes@asmah2107 A simple solution to this could be adding a retries for cache udpate failures. For a more scalable approach, we could introduce a queue to buffer the write requests and wait for Redis update to ack before dequeuing.
@asmah2107 Bonus- how does git rebase compare to these? Git resolves conflicts manually. When two branches change the same line, it halts the merge and asks the user to choose. You edit, stage, commit—then sync. All these solutions solve the same problem but in different ways.
@asmah2107 Another popular option is CRDT which is suitable for distributed systems. CRDTs resolve conflicts on the client side. Server only syncs the operations. Since merges are deterministic it leads to consistent state. This makes it offline friendly & also scalable.