And: Are tech companies cheap? + Shame on Meta!
Thursday. Welcome to the first day of Q4! Earnings season is just around the corner, and the IPO window is staring down the holiday . Let's see who gets out in time. Today, we're talking about the drama consuming Silicon Valley (and why old-school venture rules existed!), the value of tech companies, and we'll close with a few positive notes on commercializing space. Your friend, Alex.
- 📈 Trending Up:OpenAI hacking governments by accident ...Micron's business ...4x markups ...protecting humans from bots ...government borrowing costs ... performative politicking ...fuel prices ...hearing glasses ? ...
- 📉 Trending Down: Shock ... rule of law ... not bombing Iran ... paying your own way ... the free press ...
Shame on Meta: After trying to argue that its CEO's compensation should count as a tax-favorable R&D expense and a resulting argument with the IRS, Meta is claiming its data center investments are also R&D costs instead of normal, non-tax-friendly spend. As a result, Meta is saving billions.
- Come the fuck on, Meta. Other tech companies aren't taking as aggressive a tax-prep route, to their credit. I know the IRS has been gutted and the current administration is corporate-friendly, but what the hell.
Notes from the AI trenches: Google showed off benchmarks for its upcoming Gemini 4 Argon model, which appears competitive with offerings from Anthropic and OpenAI after spending months languishing behind the frontier. Google is testing the model with cybersecurity partners, though its own staff is less than impressed with Argon outside testing environments.
- OpenAI's Greg Brockman is shuttering his personal donations to the Leading the Future political action committee (PAC) after his initial $25 million gift. Employees had expressed discontent that the PAC didn't reflect OpenAI's values; Brockman said the group had become a "distraction ."
- Meta CEO Mark Zuckberg and House Speaker Mike Johnson were the seed from which the White House AI plan grew ; other AI CEOs pressed Anthropic's Dario Amodei during the confab about "why he was being so extreme in public about the capabilities of AI models and risks they posed,"the WSJ reports , to which Amodei said, "it is important to be honest with the public about model capabilities and not play down risks," the paper writes.
- OpenAI accused Moonshot , the Chinese AI lab behind the Kimi family of AI models, of aggressively distilling its own work over the summer. The American company detailed its work to limit other labs' ability to extract intelligence from its models; some AI leaders view such activity as mere competition, others less charitably. Closed models have aclear intelligence edge over their open rivals at present, perhaps because distillation is a closing door.
Silicon Valley is melting down on X
I love mess. The often buttoned-up world of startups and venture capital blew its top yesterday after a CEO accused a board observer of deception while firing him. Publicly. The rival company and its investors fired back. Ironically, some of the loudest screeches came from an investor in the original company. Let me explain.
Matan Grinberg is the CEO of Factory, an AI startup that offers agents for software development. Claiming 100x revenue growth in the last year. Grinberg wrote on X yesterday that he was terminating his board observer, Christopher Degnan. Degnan, known for being the first sales hire at Snowflake, told Factory that he was also talking to Cognition, a Factory AI rival, but that it wasn't a serious commitment.
Later, Degnan "called [Grinberg] to say that the conversation [with Cognition] that was initially described as casual and one-off was actually formal and recurring." Degnan got the boot from Factory as he announced that he was taking on the Chief Revenue Officer role at Cognition.
Now, shuffling executives is not new. Stating publicly that you are firing an advisor over ethical missteps is. Enter the pushback. After Grinberg made his case, Khosla investor and founder Vinod Khosla, whose firm has invested in Cognition, went nuclear:
Grinberg replied that he would "rather not post email receipts, but [could assure Vinod] that [his accusations are] factually incorrect." (Degnan came to his own defense, of course.)
Now things get wild. Khosla was also a named investor in Factory's recent $200 million, $5 billion valuation round announced in mid-September. Of 2026. Yes, Khosla's founder is calling a Khosla portfolio company and CEO struggling, unethical, and desperate. Not good!
Vinod's reputation for being, shall we say, spiky came to the fore with CEOs like Amplitude's Spenser Skates (one of my favorite tech execs) saying that the saga was "another great reason not to raise from Vinod," who he described as one "of the biggest bullies of founders in the VC ecosystem." Menlo's Matt Murphy weighed in on Factory's side.
Subscribe to keep reading #
Everything below is for paid subscribers