Good morning,

I have a confession to make.

I didn't feel like publishing today's newsletter. In fact, I seriously contemplated just quitting this entire project.

Cleaned out the subscriber list - a quarter of the subscribers weren't actually opening the email. I read that it's normal but still. Kind of feels silly to me. So they are gone and I kind of feel deflated about it.

Might have well have kicked me in the nuts.

I see other newsletters booming with AI-generated content and lots of Facebook marketing. While this one is mostly hand-written and I'm just hoping to spread organically.

It's like a small woodcarving shop having to compete with IKEA.

Every week i'm burning money, time and tokens.

But enough complaining.

The actual reason that the impulse to quit is that high is that I see people "building stuff" with AI. They devote their time to building new projects , learning about these models and applying them.

But not everybody is building.

I think you can grossly put people in 4 categories.

Those are the people I mentioned. Burning the midnight oil, running projects, building agent swarms. The OpenClawers.

At a certain point I was here. It goes so fast that I sometimes want to forget about AI , put my head in the sand and hope smarter people than me will figure things out and tell me my job is safe. Reality is that by not building your job won't be safe.

Yes, i've met them. Yesterday someone told me he's still coding without any AI assistance, because of the "craft" and "intellectual challenge". This person , just like me with my woodworking shop mentality will be made obsolete soon.

I think this segment is larger than we all think. Head in the sand and not interested in learning. These people will be the first ones to get laid off and ironically also the first ones on the barricades trying to burn some tires in front of their office.

So, yes.

Should i quit this token-burning mess of a newsletter and spend the time building instead? Should I keep shouting and trying to make the unaware wake up ?

Still on the fence to be honest. Guess you'll find out next week :)

Welcome to the Blacklynx Brief

DeepSeek V4 launches with cheap open-source frontier model
Chinese lab DeepSeek released preview versions of V4 with a one-million-token context window, pricing input/output at $1.74/$3.48 per million tokens -- a fraction of GPT-5.5's $5/$30 rate. Early external tests place V4 Pro near the top of open models, and Huawei confirmed its Ascend chips can run the model, demonstrating a fully functional AI stack operating outside U.S. chip supply chains. (DeepSeek V4 technical report)

Musk's $130 billion trial against OpenAI begins
Elon Musk took the stand in federal court as opening statements began in his $130 billion lawsuit against OpenAI, accusing CEO Sam Altman of betraying the organization's nonprofit mission to pursue a for-profit agenda. Musk's suit seeks the ouster of Altman and President Greg Brockman from the board and a forced return to nonprofit governance. OpenAI's attorneys called the suit "sour grapes," while Microsoft's team noted Musk did not object to OpenAI's evolving structure until xAI became a direct competitor. (CNN)

OpenAI and Microsoft rework partnership, kill AGI clause
OpenAI and Microsoft overhauled their foundational agreement, ending Microsoft's exclusivity over OpenAI's intellectual property and removing the clause that had restricted certain rights until AGI was declared. OpenAI can now deploy on rival cloud platforms, including Amazon Bedrock, while Microsoft retains Azure-first launch access through 2032 and stops paying revenue share, holding equity instead. The deal simultaneously resolved Microsoft's reported threat of legal action over OpenAI's $50 billion agreement with Amazon. (OpenAI)

Google signs classified Pentagon AI deal despite 600-plus employee protest
Google finalized a classified AI contract with the Pentagon, opening its models to use for "any lawful government purpose," despite more than 600 employees sending CEO Sundar Pichai a letter urging him to refuse classified military work. The deal follows similar Pentagon agreements signed by OpenAI and xAI last month. Anthropic is currently in court fighting a government blacklisting after declining to remove its safety guardrails as a condition for access. (The Information)

White House accuses Chinese labs of 'industrial-scale' AI theft
The White House published a formal memo accusing Chinese firms of running "industrial-scale" distillation campaigns against U.S. frontier AI models, escalating what had been a corporate dispute into an official national security matter. The memo, designated NSTM-4, follows Anthropic's February accusations against DeepSeek, Moonshot, and MiniMax. A House Foreign Affairs bill that cleared its first committee vote this week would direct the administration to sanction any company found to have stolen frontier AI intellectual property. (White House NSTM-4)

AlphaGo creator David Silver launches $1.1 billion 'Ineffable Intelligence' lab
David Silver, who led DeepMind's reinforcement learning team for a decade and built AlphaGo, AlphaZero, and AlphaProof, launched Ineffable Intelligence in London with $1.1 billion in seed funding at a $5.1 billion valuation -- Europe's largest seed round on record. The lab's goal is an AI that learns purely from experience in simulation, skipping pre-training on human-generated data entirely. (Ineffable Intelligence)

That’s it for us this week. Please like and subscribe 🙂

The answer : The Cure - A Forest