1.
[AINews] AMD buys Taalas
In The Custom ASIC Thesis we said Taalas was worth paying attention to, and in the Inference Inflection we said everything would go vertical. Our Baseten episode had some skeptical counterpoints against etched LLMs, not just custom ASICs, but clearly Lisa Su disagrees for now. Congrats! AI News for 8/5/2026-8/6/2026. We checked 12 subreddits, 544 Twitters and no further Discords. AINews’ website lets you search all past issues. As a reminder, AINews is now a section of Latent Space.
2.
Mashable will send the best deals directly to your phone daily — how to sign up for free
They're constantly scanning top retailers like Amazon, Walmart, and Best Buy to find limited-time discounts on products they'd actually recommend. You can expect to get a Deal of the Day most days, but during major shopping events like Prime Day and Black Friday, they might send multiple deals per day if they find enough good offers.
The text group is designed to bring you the best deals on products like MacBooks, AirPods, and items under $50. They're not just sending any old discount, but rather carefully curated deals that are worth your time and money.
Signing up is quick and easy, and you can do it right from their website. They're not going to sell your number or spam you with unwanted messages – they just want to help you save some cash.
I think it's a great idea, and I'm actually considering signing up myself. I'd rather get these deals directly to my phone than spend hours searching for them online.
3.
Scientists Make First Viruses Designed By AI
I’ve been thinking about this AI‑crafted phage work—basically, researchers fed a genome‑language model the code of a bacteriophage, let it remix the sequence, and then built the virus in the lab. The result was a cocktail that wiped out E. coli strains that even natural phages couldn’t touch, showing how quickly we can tailor genomes to a specific bacterial foe.
The team is quick to flag the flip side: designing whole viral genomes with generative AI opens a new biosafety conversation. They’re urging anyone tinkering with these tools to keep safety and security experts in the loop from day one.
What’s striking is that the real control point isn’t the AI itself but the downstream steps—who gets access to the DNA, how synthesis is screened, and the lab practices that contain the work. A layered approach, from model access to synthesis oversight, seems to be the consensus for keeping the promise while curbing the risk.
4.
Sarvam is India’s AI moment. What comes next counts.
The country’s poster boy for Indic AI models, Sarvam AI, recently unveiled plans for a trillion-plus parameter model even as it launched an India-hosted inference platform, opened a San Francisco research office and has reportedly raised another $75 million in a funding round led by Nvidia Corp., just months after securing $234 million. The Bengaluru-based AI startup plans to use the money to fund its next-generation AI models for applications such as AI agents, software development and cybersecurity, while expanding the computing capacity needed to deploy them across industries.
5.
Inside the 1-Bit LLM: How Bonsai Fits a 27B Model on a Phone
In July 2026, PrismML shipped a 27-billion-parameter language model, built on Alibaba's Qwen3.6-27B, that runs in a browser tab on WebGPU and, separately, on a phone. The full-precision version of that model is 54 gigabytes. The one that fits in your pocket is under 4. This post is about how that gap gets closed, what the arithmetic behind it actually says, and where the honest line sits between what PrismML has disclosed about their method and what remains their own.
6.
How to Build Your First AI Agent: A Beginner's Guide
Create an empty folder on your computer. Drop three messy notes inside it: a meeting transcript, a project update, and a page of customer feedback. Then give an AI one clear finish line: Read every file. Turn them into a clean daily brief. Mention the source behind every important point. Run a check on the final document. Fix anything that fails. Stop only when the check passes. A few minutes later, the folder contains a finished brief. You did not copy anything from a chat. You did not ask for a second draft. You did not remind the model which file it had forgotten.
7.
ThursdAI - Aug 06 - Google shakeup, Details on OpenAI hack, 2 new agent harnesses, 4 video models (1 Open) and 3 guest segments
This week we saw a major shakeup at Google, with the departure of long time folks like Jeff Dean, and Oriol Vinyals, Demis stepping down from leading DeepMind, and the delayed release of the improved Gemini. While this was a big deal, it’s not the only one worth covering as the details of the OpenAI hack (and 2 new ones from Meta and Anthropic) came to light, as well as new details from the UK AI Security Institute. As mentioned on the show, CoreWeave is coming to SF for Fully Connected, our premier 2000 person AI event.
8.
💧⚡🖥️ A deep-dive on data centers: My interview with AI researcher Andy Masley
My fellow pro-growth/progress/abundance Up Wingers in America and around the world: With the massive buildout of data centers has come an equally massive backlash. They’ve been blamed for higher electricity bills, taking up too much water, and pushing out farmland. All this has led to a public outcry and efforts against their construction.
9.
Ukraine, Iran, and America’s Arms Overreach (Robert Wright & Jennifer Kavanagh)
This episode includes an Overtime segment that’s available to paid subscribers. If you’re not a paid subscriber, we of course encourage you to become one. If you are a paid subscriber, you can access the complete conversation either via the audio and video on this post or by setting up the paid-subscriber podcast feed, which includes all exclusive audio content. To set that up, simply grab the RSS feed from this page (by first clicking the “...” icon on the audio player) and paste it into your favorite podcast app.
10.
ChatGPT Rated My LinkedIn Photo in 60 Seconds (And Told Me How To Fix It!)
I thought ChatGPT was about to critique my face. Honestly, I almost didn’t upload the photo. But less than a minute later, I had something much more useful: a detailed report on how my LinkedIn picture comes across to recruiters, clients, and colleagues. No comments about attractiveness. No vague advice to “look more professional.” Just practical feedback on the crop, lighting, background, clothing contrast, eye contact, and overall credibility of the photo. Tiny details. The kind I had looked at a hundred times without actually seeing.