Are We All Retiring in 2029?

Good morning.

As this newsletter steadily grows, we have the occasional critic showing up in our feed, which we obviously welcome.

So far, I’ve been accused of being an Elon Musk fanboy, of not understanding the limitations of Large Language models, and of pushing the narrative that AI is going to replace all jobs. Which, according to one pretty vehement critic has zero possibility of happening.

But the more I read about it , the more I’m convinced “knowledge work” is going the way of the dodo.

You just have to pay attention to what’s happening.

Every week, there’s a snippet or two of news that makes me frown a little bit.

For example, this week the annual Bilderberg summit takes place, which involves discussions between “thought leaders” from Europe and North-America on pressing issues. According to conspiracy theorists, these are the "illuminati,” but let’s not mention that.

Anyway, guess what is hot on the agenda: AI and the threat of AI to our jobs. The CEOs of Google Deepmind, Anthropic, Mistral, etc. have been invited. Make of that what you will.

Then earlier this week there was this somewhat strange article from Anthropic’s chief of staff Avital Balwit called “My Last Five Years of Work”.

She is witnessing the change first hand and points at the denial many people are showing.

But ok, she’s a 25 year old copywriter, you might say. What does she know ?

Then there was this other paper - completely different - from an OpenAI researcher Leopold Aschenbrenner. Well, former OpenAI, as he was fired for having safety concerns.

He is expecting Artificial General Intelligence to arrive by 2027 and that hundreds of millions of AGI’s would compress many decades of progress into one year.

Oh and one last thing: the IMF, the freaking International Monetary Fund, is publishing articles on AGI.

Do with that information what you want , but ignore at your own peril I would say.

Welcome to the Blacklynx Brief.

AI NEWS
Death to the Superbugs

• Researchers used AI to predict nearly one million new antibiotics hidden in microbes worldwide, uncovering potential treatments for bacteria and superbugs. By analyzing data from over 100,000 genomes, the AI identified parts that might produce antibiotic compounds, with lab tests confirming 79 out of 100 candidates as potential antibiotics. This breakthrough, called the largest antibiotic discovery ever, could rapidly address antibiotic-resistant bacteria, offering new hope against superbugs.

• OpenAI board members Bret Taylor and Larry Summers defended CEO Sam Altman against claims by former board members Helen Toner and Tasha McCauley. They cited an external review that found no evidence Altman was fired due to safety or financial issues and emphasized OpenAI's regulatory efforts. The public dispute highlights ongoing tension, with Toner accusing Altman of lying and manipulating board members.

• AI startup Perplexity has launched Pages, a feature for creating customizable, visually appealing AI-generated articles, guides, and reports. Pages use Perplexity's AI search models to generate structured reports with citations and allow users to tailor writing styles and modify structures. The tool also adds visuals and enables content sharing and publishing.

• Nvidia CEO Jensen Huang announced new AI initiatives at the Computex conference, including next-gen 'Rubin' chips, an AI gaming assistant, and tools for creating lifelike avatars. The 'Rubin' platform, expected in 2026, aims to spark a "new industrial revolution," while Project G-Assist offers personalized gaming help. Additionally, the ACE suite simplifies the creation of digital avatars for uses like customer service and healthcare. Not everyone is happy with what NVIDIA is doing.
  

• As mentioned above, former OpenAI researcher Leopold Aschenbrenner predicts rapid and volatile AI progress, comparing it to the development leap from GPT-2 to GPT-4 by 2027, which he believes will lead to AGI. He suggests AGI will greatly accelerate technological advancements, raising economic and military stakes, and views it as a national security issue needing urgent attention. Aschenbrenner's comments on his dismissal for raising AI safety concerns add to the ongoing controversy at OpenAI.

• Leaked internal emails reveal Elon Musk instructed Nvidia to prioritize AI processors for X and xAI, delaying Tesla’s $500M order. Musk redirected 12,000 H100 GPUs to X in December, contradicting Tesla's public plans to expand AI hardware purchases. While Musk claims Tesla lacks installation capacity for the chips, this move fuels criticism amid Tesla's ambitious self-driving targets.

Quickfire News

• OpenAI is relaunching its robotics team to develop models for other companies' robots, after discontinuing its own robotics division in 2020.

• You.com launched Custom Assistants, allowing users to create personalized AI bots using large language models like GPT-4o and Claude 3.

• Maven AGI announced a $28M funding raise, highlighting its customer support AI that resolves over 93% of issues and reduces support costs by 81%.

• Perplexity is reportedly in talks to raise $250M at a $3B valuation, with the company generating over $20M in annual subscription revenue.

• Anthropic has enabled Claude to interact with tools for all users, allowing for more complex tasks and enhanced capabilities.

• ElevenLabs launched Text to Sound Effects, a new tool that generates sound effects, short instrumental tracks, and soundscapes from text prompts.

• Anthropic Chief of Staff Avital Balwit wrote an article titled ‘My Last Five Years of Work,’ stating that AI advancements in the coming years will 'end employment' as we know it.

• Top AI leaders are attending the Bilderberg Meeting this week, with CEOs from Google DeepMind, Microsoft AI, Anthropic, and Mistral reportedly discussing AI, geopolitics, and global economic challenges.

• Sam Altman participated in an interview at the AI for Good Global Summit, discussing GPT-5, the Scarlett Johansson voice controversy, and Helen Toner’s recent commentary.

• Intel introduced new AI advancements at Computex: Xeon 6 processors, Gaudi accelerators, and Lunar Lake architecture.

• Stability AI announced the open release of Stable Diffusion 3-Medium for June 12, with a smaller 2B parameter model.

• AMD unveiled next-gen MI325x AI chips at Computex, with MI350 planned for 2025 and MI400 series for 2026.

• Microsoft announced a $3.2B investment in Sweden’s AI and cloud infrastructure to train 250,000 workers by 2027 and enhance data center capacity.

• Google addressed AI Overview issues in a blog, highlighting new detection mechanisms and ongoing accuracy efforts.

• Adobe is facing backlash from the estate of Ansel Adams for including AI-generated images labeled “Ansel Adams-style” in Adobe Stock, violating the platform’s ToS.

• Zoom founder Eric Yuan mentioned in an interview with The Verge that 'digital twin' AI avatars will eventually attend Zoom meetings and make decisions on behalf of users.

• OpenAI's latest mobile app update for ChatGPT includes a new feature called 'Background Conversations', allowing users to continue their conversations even after exiting the app.

• Pika Labs raised $80M in Series B funding for its text-to-video platform, now valued at $470M, with investors like Spark Capital and actor-musician Jared Leto.

• Actor and entrepreneur Ashton Kutcher is currently testing OpenAI’s Sora and mentioned that some of its outputs are already good enough for use in movies and TV shows.

• TSMC Chairman C.C. Wei stated that the proposal by OpenAI's Sam Altman to create a global network of AI chip factories was too ambitious, and the company couldn't commit to the required number of fabs.

• xAI has agreed to build a 'gigafactory of compute' supercomputer in Memphis, TN, aiming for completion by late 2025.

Closing Thoughts

That’s it for us this week.

If you want a version of this newsletter with pictures etc. Be sure to subscribe below:

https://www.blacklynxbrief.com/subscribe