Naveed Ahmad

I Work in Hollywood. Everyone Who Used to Make TV Is Now Secretly Training AI

People who had previously felt paralyzed by their NDA’s began to talk. Helena, the conflict-avoidant moderator of the Mercor subreddit, worked overtime deleting furious rants from aggrieved workers who delighted in dropping names of the “secret projects”—something explicitly banned by the non-disclosure agreement every tasker must sign before being hired as an Independent Contractor. Elsewhere,…

Read More

Bad news for internet users

Internet users in Pakistan may experience slower connectivity in the coming days, particularly those relying on services from Pakistan Telecommunication Company Limited. According to an official notice issued by PTCL, internet speeds across the country could be affected from May 11 to May 18 due to technical repair work on a submarine cable. During this…

Read More

Sakana AI and NVIDIA Introduce TwELL with CUDA Kernels for 20.5% Inference and 21.9% Training Speedup in LLMs

Scaling large language models (LLMs) is expensive. Every token processed during inference and every gradient computed during training flows through feedforward layers that account for over two-thirds of model parameters and more than 80% of total FLOPs in larger models. A team researchers from Sakana AI and NVIDIA have worked on a new research that…

Read More

IT exports projected to reach $4bn during current fiscal year, PM Shehbaz told – Business

Prime Minister Shehbaz Sharif was told on Monday that Pakistan’s information technology exports were expected to reach $4.5 billion to $4.6bn, according to the Prime Minister’s Office (PMO). The development came as the premier chaired the meeting to review the affairs of the Ministry of Information Technology and Telecommunication. During the meeting, the premier was…

Read More

A Coding Implementation to Build Agent-Native Memory Infrastructure with Memori for Persistent Multi-User and Multi-Session LLM Applications

banner(“Part 5 — Streaming”) mem.attribution(entity_id=”[email protected]”, process_id=”personal-assistant”) stream = client.chat.completions.create( model=MODEL, messages=[{“role”: “user”, “content”: “In two sentences, what do you remember about me?”}], stream=True, ) print(“[stream] “, end=””) for chunk in stream: d = chunk.choices[0].delta.content if d: print(d, end=””, flush=True) print(); time.sleep(WRITE_DELAY) banner(“Part 6 — Async LLM calls”) async def async_demo(): r = await async_client.chat.completions.create( model=MODEL,…

Read More

Whitehat Returns $190K to Renegade After Hacking Them

The team behind the Renegade.fi protocol said a whitehat hacker returned about $190,000 after exploiting one of its Arbitrum-based decentralized dark pools and later complying with instructions in an onchain message to return 90% of the funds. Renegade confirmed the return of funds on Sunday after blockchain analytics platform Blockaid flagged the $209,000 exploit at…

Read More