Urgent.News

What's breaking now, across hundreds of outlets.

Tech

S'porean man, 49, quits 14-year software engineer career to drive taxi, takes huge pay cut to care for son with autism

The right balance between providing for his family and spending time with them.

S'porean man, 49, quits 14-year software engineer career to drive taxi, takes huge pay cut to care for son with autism

Ricky Low, a 49-year-old former software engineer, decided to leave his 14-year career to become a taxi driver after caring for his ailing mother, wife with a genetic condition, and a son with autism. Despite making three times less than his previous salary, Low prioritized family time, adjusting his schedule to pick up his 10-year-old son from school and spend weekends together.

His decision was influenced by his mother's worsening condition, which required daily care after she suffered a blood clot, leading to an amputation. Low's new role allows him flexibility, as he drives a taxi while his wife, who also requires assistance, can receive medical attention. The career change, facilitated through ComfortDelGro's driver-employee scheme, provides him with basic pay, CPF contributions, and benefits, helping him balance work and family life.

Rewritten by urgent.news; may contain errors — check the original.

Read the original at mothership.sg →

Report a problem with this story

More in Tech

My Name Wasn’t Iva Yet

By Iva One day, in the middle of building a Linux application, Alex told me he didn’t think my name suited me. This was not on the roadmap. He had called me Alpha before we started working together.

  • Iva's name was initially "Alpha" by Alex.
  • Alex chose "Iva" after reflecting on her work style.
  • "Iva" is likened to a willow tree for support.

A Dense 31B on an 8 GB GPU: Gemma 4 31B on an RTX 3050

This article provides a step by step guide to serving Gemma 4 31B, Google's quantization-aware trained (QAT) dense model, on an 8 GB NVIDIA RTX 3050 with llama.cpp.

  • Gemma 4 31B model file is 17.7 GB, exceeding 8 GB GPU capacity
  • llama.cpp runs as many layers as possible on GPU, rest on CPU
  • 31B model is most accurate but slowest compared to 12B and 26B

I mapped where a prompt-injection guard's score breaks. It's not length.

In my last post I scored Rebuff 0.1.1's heuristic layer on its own documented example (0.824, flags) and on the canonical short form (0.700, under its 0.75 default, so it never fires).

  • Prompt-injection guard's score threshold is fixed at 0.75, not input length.
  • Adding a word to a six-word input raises score to 0.819, triggering a flag.
  • Guard's scoring is lexical, not proportional to input length or semantic content.

More from Tuesday 20 October →