Thursday energy is here - let's make the most of it!
Geek Gossip
#XRevenueSharing - X (Twitter) came out with an update stating that content creators will now earn revenue based on engagement of readers within their own country / region - only to immediately put it on hold after Elon Musk (who owns X) disapproved of it. One would think they run it by their boss for approval before making announcements public, but we guess not.
#GitHub - WATCH OUT. Github has updated their terms and will now use your codebases to train their AI models. It is opt-in by default and you have to explicitly opt out.
#MetaLayoffs - Meta has reportedly begun laying off employees as it focuses more of its cash on building out datacenters, training its own large language models, and recruiting talent for AI. CEO Mark Zuckerberg said this was the year Meta would begin “flattening teams” in a post-earnings note. The company is also designing its own custom chips for GenAI workloads which it plans to build over the next two years.
Latest Updates
Wikipedia Bans AI-Generated Text In Articles, With Exceptions
Wikipedia has officially banned the use of large language models (LLMs) for generating or rewriting article content, except for refining writing and translation assistance, where editors still need to check for accuracy and errors.
Reddit Introduces New Measures To Distinguish Between Humans And Bots
Reddit is taking steps to combat the increasing presence of bots on its platform. The company will start labeling non-human accounts with a [App] designation, allowing users to know when they're interacting with a machine rather than a person.
Samsung Unveils Galaxy A57 5g And A37 5g
The new Galaxy A57 5G and Galaxy A37 5G bring Samsung’s newest mobile innovations - including enriched Awesome Intelligence - to more users worldwide. Galaxy A 57 5G, featuring a slimmer, more refined design and upgraded AI-powered capabilities, is the most powerful Galaxy A series device yet.
Apple Enhances App Store Connect With More Than 100 New Metrics
Apple's platform for developers to track app performance and monetization, now includes more detailed data and tools to help them understand user behavior and compare with industry peers.
Social Media Giants Held Liable For Mental Health Harms
Jury finds Instagram and YouTube failed to warn users about the risks of using their products. The jury ordered both companies to pay a total of $3 million in compensatory damages. Jurors found that punitive damages were warranted, and they’ll deliberate further on the appropriate amount to award.
Changelog
Grapevine - Launches WhatsApp-based AI talent agent "tal" to revolutionize job discovery and career guidance for Indian professionals.
Google Meet - New safeguarded guest admit flow with two separate queues for meeting requests.
Firefox - Free built-in VPN feature for secure browsing and location protection.
Claude Code - Introduces Auto Mode, a research preview that allows Claude to make permission decisions on your behalf with safeguards in place.
Lyria 3 Pro - Google launches new version of their AI music generating service.
Tech Talk Thursdays
Next.js 16.2 introduces major updates designed to optimize the framework for AI agents like Claude and Cursor. Key features include a bundled local documentation file that helps agents avoid hallucinations, and "Next Browser," an experimental CLI that allows AI to visually inspect the component tree and take screenshots for debugging. Performance sees a massive boost, with dev startup times increasing by up to 400%, alongside new hydration diff indicators and terminal-based server function logging.
In Case You Missed It
🎁 We now have a dedicated page to list all events. We will be updating this page EVERY WEEK. We will of course be publishing new (and coming-soon) events in this section but for a comprehensive list of events, check the link(s) below:
This link is EXCLUSIVELY accessible by subscribers of Heap Heap Hurray!
Miscellaneous
8 Design Principles for the Agentic Enterprise - From Salesforce
AI failures in deployment are often not due to the model itself, but rather the underlying architecture. A successful agentic deployment requires four architectural layers: system of engagement, system of agency, system of work, and system of data. To achieve this, companies need to design for modularity, open standards, and vendor lock-in.
Interviewing Tactics For A Post-LLM World
Traditional take-home assignments are becoming obsolete as LLMs make it difficult to distinguish between human and machine intelligence. Companies are changing recruitment processes, but the answer lies in designing interviews that leverage LLMs, focusing on contextual judgment, critical thinking, and real-world experience to evaluate genuine expertise.
Attention Subscribers
Every two weeks, on Friday at 7 PM IST, we distribute a “BONUS EDITION” featuring highly valuable links. These special editions are exclusively sent to our most dedicated and engaged subscribers. To ensure you receive these editions, please take the following actions:
1. Regularly open our emails. We track email open rates to identify our most active subscribers.
2. Engage with the links in our newsletter. While we provide concise summaries, your engagement is equally important.
Your active participation ensures you continue to receive these exclusive benefits. Thank you for your ongoing support and engagement.
If you are seeing the bonus edition posts in your SPAM folder, please do us a favour and mark it as NOT SPAM.