DS

Still learning ...

[DS @dsing42]

@michaeljburry Zuck = One hit wonder

[DS @dsing42]

@michaeljburry First xAI and now Meta. Both spent 10s of billions on data centers and NVDA chips.

Found out they have little use for all this compute. No one uses their AI models.

Now selling it to get some return.

Market has not caught on to the compute overcapacity yet. Give it a month.

[DS @dsing42]

@jasonlk Explained by this maybe?

92 P/E

[DS @dsing42]

@PodcastAlphaX @elonmusk @chamath FALSE!

The cursor coding model iuses a Chinese open source model as the base.

[DS @dsing42]

This is utterly ridiculous.

DeepSeek, Qwen and https://t.co/27QC4zfo4P have published AI research papers on novel techniques they use to make their model training and inference more efficient and effective. That is leading edge work.

This post is just pushing investor narrative when it is clear that models are becoming a commodity.

Chat, text, video, voice models are already commodity. Only coding and agentic use is differentiated, but that too will be commodity by later this year (look at MiniMax M3 and GLM 5.2).

[DS @dsing42]

@mehtadata_ Every tool you cut removes a seam, and seams are where these stacks break. The GTM engineer role works because one person owning the whole flow beats five owning fragments of it.

[DS @dsing42]

Everyone has the tools now. Ship something with them before the feeling of productivity becomes the whole job.

[DS @dsing42]

@Tim_Denning Where's the line between leading and performing it online? The best leaders I've worked under were nearly invisible. Too busy doing the work to narrate it.

[DS @dsing42]

@davidsenra Jobs timed that IPO to the story, not the numbers. A hit in theaters is the narrative markets overpay for. Most founders raise on metrics and miss the window entirely.

[DS @dsing42]

@gavinzaentz A funding alert you act on three days late is just news, not a buying signal. The window is the whole value.

[DS @dsing42]

Volume was never the constraint. Knowing which accounts are actually in-market this quarter is the hard part, and no amount of send-speed fixes a bad target list.

[Alexa | Startup founder @alexabelonix]

If your AI SDR is proud of sending 5,000 emails, I’m already nervous.

The market does not need more automated enthusiasm.

TechCrunch covered Actively AI saying classic AI SDRs failed because they played the “pure volume” game.

That’s exactly the trap.

The better sales AI does this:

reads CRM
finds patterns in won deals
checks account signals
ranks high-value prospects
explains why they matter
suggests the next best action

The dumb version writes more emails.

The smart version tells the founder which 12 people are worth human energy today.

[DS @dsing42]

One customer once ate 40% of our support hours and still churned inside a year. Closing them cost us two good accounts we were too busy to serve. Winning the wrong-fit deal is the most expensive thing a sales team can do.

[DS @dsing42]

I got my biggest raise the year I stopped trying to look busy and started saying no to the work that didn't matter.

[DS @dsing42]

Most sales forecasts are graveyards. Deals nobody wants to call dead, still sitting there taking up space and hope.

[DS @dsing42]

Why do we promote the best rep to manager, then act surprised when we lose a great seller and gain a mediocre coach?

[DS @dsing42]

@IAmAaronWill Shrink the bet until a no costs you nothing. Most paralysis is just overpricing the downside of being told no.

[DS @dsing42]

@geoffwoo Those four tabs and the manager sign-off were never about the refund. They were blame insurance. The agent quietly deletes the step that only protected the org, not the buyer.

[DS @dsing42]

@danmartell The fear never fully leaves. At 3Com we bet the company on broadband while most homes were still on dial-up. Felt reckless for years. Being early just looks like being wrong until the day it doesn't.

[DS @dsing42]

@jasonlk @MonacoGTM A tool that generates its own pipeline has already run the only experiment that counts. Everything else is just a deck.

[DS @dsing42]

@remoteoliver Which of the two do you quietly avoid looking at? That's the half capping you. The skill you dodge is never the one you've actually fixed.