CO

Building @staypace

[CO @cetricokola]

Jobs demonstrated a profound truth:

The best companies don’t just sell products; they sell experiences.

His approach was straightforward:

1. Concentrate on a select few products.

2. Strive for perfection.

3. Embrace innovative thinking.

[Mark Karpelès @MagicalTux]

Revolut customers were targeted in a fraudulent emergency data request sent via an email at a "government agency." https://t.co/sS2sbwAt8r

[CO @cetricokola]

Check out my latest article: The Revolut Incident Is a Reminder: A Verified Email Is Not a Verified Authority https://t.co/KXOqs9ienn via @LinkedIn

[GLOBALITE 👾 @Globalite27]

Kapıştırılan boğalar kafa kafaya çarpıştı!

[CO @cetricokola]

Check out my latest article: The Next Fintech Battle in Kenya Is Not Payments — It Is Financial Infrastructure https://t.co/RJj3cMrvwL via @LinkedIn

[CO @cetricokola]

Check out my latest article: Why Kenyan SMEs Need Financial Operating Systems, Not Just Payment Gateways https://t.co/Kkv9LYEFfX via @LinkedIn

[Geno Insights | Analysis @GenoInsightsKe]

Tough lessons learnt

[CO @cetricokola]

Check out my latest article: Idempotency in Fintech: The Small Engineering Detail That Can Prevent Very Expensive Mistakes https://t.co/xinUTmTlRO via @LinkedIn

[Eric @amerix]

Every season ends.

The difficult one you are in will end.

The good one you are in will also end.

Endure the hard ones without despair.

Appreciate the good ones without clinging.

That is wisdom.

[CO @cetricokola]

Check out my latest article: From Passwords to Biometrics: How Authentication Is Changing Finance https://t.co/HklDlNUMMn via @LinkedIn

[CO @cetricokola]

Check out my latest article: WhatsApp Business API Pricing Is Changing: What the New Charges Mean for Kenyan Businesses https://t.co/JXFVn4E4YZ via @LinkedIn

[evansoo 🇰🇪 🇨🇦 @elangat016]

TUJIFUNZE KISWAHILI:
1. Charger - kimemeshi
2. Remote - kitenzambali
3. Password - nywila
4. Akala (bladder shoes) - kirikiri
5. Passion fruit - karakara
6. Kangaroo - bukunyika
7. Fridge - jokofu
8. Juice - sharubati
9. Chips - vibanzi
10. PHD - uzamifu
11. Masters degree - uzamili
12. Degree - shahada
13. Diploma - stashahada
14. Certificate - astashahada
15. Keyboard - kicharazio
16. Scanner - mdaki
17. Flash disk - diski mweko
18. Mouse (for a computer) - kiteuzi
19. Floppy disk - diski tepetevu
20. Computer virus - mtaliga
21. Distillation - ukenekaji
22. Evaporation - mvukizo
23. Synthesis - uoanishaji
24. Oesophagus - umio
25. Greenhouse - kivungulio
26. Femur - fupaja
27. Germ cell - selizazi
28. Humus - mboji
29. Nector - mbochi/ntwe
30. Nutrients - virutubisho
31. Crystal - fuwele
32. Appetizers - vihamuzi
33. ATM - kiotomotela
34. Business card - kadikazi
35. Scratch card - kadi hela
36. Simcard - kadiwia/mkamimo
37. Memory card - kadi sakima
38. Microwave - tanuri ya miale
39. Mobile phone - rununu/rukono
40. Stapler - kibanizi
41. Laptop - kipakatalishi
42. Power saw - msumeno oto
43. Duplicating machine - kirudufu
44. Photocopier - kinukuzi
45. Cocktail party - tafrija mchapulo
46. Air conditioner - kiyoyozi
47. Calculator - kikokotozi
48. Lift - kambarau
49. Reaction - radiamali
50. Toothpick - kimbaka

[Jonathan Wilke @jonathan_wilke]

By the way, https://t.co/3tMzhIlC1s was 100% built with Grok 4.6 in Cursor.

[Shubham Mishra 🐐 @shubh6200]

how does WhatsApp actually handle it? They use two core architecture principles: Lazy Deletion and Time-To-Live (TTL) storage.

Here is the breakdown of the system design:

1. The Illusion of Lazy Deletion -
Instead of actively deleting data at the exact 24-hour mark, you handle it on the read path. When a user uploads a status, stamp it with expires_at = NOW() + 24 hours.

When friends fetch the status feed, the API filters it: WHERE expires_at > NOW(). If the time has passed, the API ignores it. To the user, it has disappeared. Physically, it still exists in the database for a bit, but it costs zero performance to "hide" it.

2. NoSQL & Native TTL -
For metadata, leverage a distributed NoSQL database like Apache Cassandra or ScyllaDB. When writing the row, set a native 24-hour TTL. The database handles the cleanup under the hood during routine background compaction, bypassing the query engine entirely.

3. Cache Eviction -
Keep active statuses in a distributed cache like Redis. Set a hard expiration on the keys (EXPIRE key 86400). Redis evicts expired keys passively when accessed or via a background sampling algorithm at O(1) efficiency.

4. DB Partitioning (The Relational Alternative) -
If you must use a relational DB like PostgreSQL, partition your tables by day (e.g., status_2026_08_20). Instead of deleting billions of individual rows, you just drop the entire daily partition table after 48 hours. Dropping a table is an instantaneous file-system operation.

5. S3 Lifecycle Policies for Media
Do not let your application code touch the deletion of images and videos. Store them in Object Storage (like Amazon S3) and apply a Lifecycle Policy to automatically purge objects older than 25 hours.

The takeaway? At scale, the best way to delete data is often not to delete it at all just hide it, and let your infrastructure clean it up asynchronously.

[Akintola Steve @Akintola_steve]

No system is immune to failure.

Hardware fails.
Networks fail.
Databases fail.
Humans make mistakes.
Software has bugs.

The goal of fault tolerance isn’t to prevent failure.

It’s to keep the system useful when failure happens.

Here’s how you actually design for it:

[Uncle Abu ™ @iamabubakhr]

Bruno Guimaraes off >> Declan Rice on

Myles Lewis Skelly off >> Zubimendi on

Martin Ødegaard off >> Eberechi Eze on

Kai Havertz off >> Viktor Gyökeres on

Calafiori off >> Piero Hincapie on

I can’t believe this is my life. Thank you Mikel Arteta for changing our poor lives!❤️

[H.O.P SME @Mapimpi_]

Niggas are out there living my dream. Mountain views, lots of trees and a farm

[CO @cetricokola]

#OkolaMbele

[CO @cetricokola]

#OkolaMbele