GoKawiil
Tech News
Cloudflare engineers found that its internal load-balancing service, Pingora Backend Router, was consuming far more memory than necessary in its pingora-ketama consistent-hashing library. By reworking the underlying algorithm with better math and Rust code, the team eliminated the excess memory usage across its global fleet.
Diogo Almeida, a former OpenAI researcher who helped invent RLHF and build ChatGPT, has released a new transformer-based model called Jev through his startup TypeSafe AI. Unlike large language models, Jev doesn't generate text—it produces calibrated probability outputs, making it faster, cheaper and immune to hallucination. Developers have shown strong demand, with the company briefly unable to keep up with API traffic.
A technical resource describes methods for improving large-scale data replication by using parallel read operations and optimized write strategies. The material is available as a downloadable guide after users register for the hosting platform's content hub.
Thousands of Barclays employees in the United Kingdom signed an open letter opposing the bank's shift from a two-day to a three-day office requirement, with senior staff expected in four days. The letter asks Barclays to compensate workers for the added commuting expenses the new policy creates.
Android 17 QPR1, rolling out to Pixel devices, reportedly changes how the OS exposes developer and USB debugging status to installed apps. This means apps can no longer easily detect that a device has developer options or debugging enabled, a check many banking and finance apps use to block access.
During the recent conflict with Iran, a US special operations analyst used an AI chatbot to help produce an intelligence report claiming a Chinese vessel was carrying nuclear weapons components. The military mobilized to intercept the ship, with armed personnel and aircraft readied, before officials discovered the report was based on a hallucinated AI output and called off the operation.
A New York Times opinion editor defended the practice of publishing op-eds under politicians' names even when the pieces are ghostwritten or assisted by staff, arguing this has long been standard practice. The comment came amid broader discussion about how AI tools are now blurring the line between human and machine-assisted writing in opinion journalism.
Newly unsealed court documents in The New York Times' lawsuit against OpenAI and Microsoft reveal internal statements warning that their AI models were creating a 'doom loop' harming the web. A Microsoft applied science director reportedly described the companies' data scraping practices as the 'largest theft of labor in human history' and a mockery of fair use. Microsoft has distanced itself from these remarks, calling them one employee's personal opinion rather than the company's official stance.
Chinese AI startup Manus is reportedly negotiating a $500 million funding round valuing it at $4 billion, according to the Wall Street Journal, after resuming operations as an independent company. Prospective backers include IDG Capital, Boyu Capital, CATL, and returning investors Tencent, HSG and Zhenfund, with the company also weighing a restructuring ahead of a possible Hong Kong IPO.
Amazon has cut the price of Bose's Ultra Open Earbuds from $299 to $199, a $100 discount that marks their lowest price yet. The clip-on, cuff-style buds sit outside the ear canal, offering situational awareness alongside fuller sound than typical open-ear designs.
Boston Longevity Week, an 11-day series of events including biomarker conferences, an AI hackathon and an investor pitch night with over 30 venture firms, culminates in a Translational Longevity Summit. The event's organizers argue the longevity field's core challenge isn't a lack of innovation or even regulation, but a failure to translate promising research into validated, clinically usable treatments.
Nvidia's Nader Khalil and Sydney Sykes will lead a session at TechCrunch Disrupt 2026, running October 13-15 in San Francisco, examining how startups should choose between proprietary and open AI models. The panel will weigh trade-offs in cost, control, and speed as both open and closed model capabilities keep advancing rapidly.