Farzad: SpaceX and Tesla Just Bet Everything on One Texas Factory

Farzad posted SpaceX and Tesla Just Bet Everything on One Texas Factory and Grok Was Called Useless. Then It Matched GPT-5.6 (Shorts) with Full video replies, argued the way to beat AI doomers is Abundance made undeniable (Tesla surviving years of media hate and “most likely” about to post a record quarter), RTd Aaron Levie on agentic workloads (swarms, computer use, APIs/MCPs, new form factors) and replied that token generation is unbelievably early, quote-tweeted operationdanish that AI’s health impact looks more positive than he expected, liked CompleteSkeptic’s Jev/RLCD release thread, RTd Elon on Starlink V3 starting this month with >100x bandwidth vs the current ~11,000-sat constellation, liked and RTd Sawyer Merritt’s Starship Flight 14 / Starlink V3 capacity stats, and said AI will drive productivity exhaustion in humans while the real fight is distributing the tech to the many instead of the few.

SpaceX/Tesla Texas Factory Short + Full video

Farzad, Sep 16 morning CT: SpaceX and Tesla Just Bet Everything on One Texas Factory (Short). Reply: Full video to YouTube The Secret Ring Under Elon Musk's Brand New Ultra Factory (id Ix6yc6R0seg). Titles only; do not invent what the video shows beyond the titles. (Same Short+Full pattern posted twice ~6 minutes apart; links below are the later pair.)

Check Short here. Full video reply here.

Grok Short + Full video

Farzad, Sep 15 evening CT: Grok Was Called Useless. Then It Matched GPT-5.6 (Short). Reply: Full video to YouTube id H0NNRZd_vK4 (same long-form id as the prior 5 AI CEOs Just Said the Same Thing Full video replies). Titles only; do not invent benchmark details beyond the Short title.

Check Short here. Full video reply here.

Abundance vs doomers + Tesla

Farzad, Sep 15 afternoon CT: the only way to defeat the AI Doomers is delivering Abundance as fast as possible; make the technology UNDENIABLE; Tesla survived 2–3 years of intense media hate and is most likely about to post a record quarter; don’t fight the narrative - prove them wrong with results. Framing and “record quarter” line are Farzad’s (not a filed earnings print in the tweet).

Check the post here.

Levie agent workloads RT + Farzad reply

Aaron Levie (@levie), Sep 15: we probably need to update our sense of what’s coming on agentic workloads - agent swarms, better computer use, next wave of APIs and MCPs, new form factors like Muse or Instinct, vertical enterprise agents, background workflow agents (per levie). Farzad RTd. Farzad reply to levie: token generation is unbelievably early; what happens when agents get extremely good at creative screen work like editing a video? Levie lines are his; the reply is Farzad’s.

Check Farzad RT here. Levie here. Farzad reply here.

AI health QT + distribute-to-the-many

Farzad, Sep 15 early afternoon CT, quote-tweeting @operationdanish (“Thanks Dario”): this is significantly more positive than he expected; a large % of the population - even with AI doomerism - sees the positive impact of AI on health; keep delivering and adoption becomes unstoppable. Separate post same window: we’re WAY past the point that AI will kill us all because everyone is aware of that risk; now it’s about ensuring the technology is distributed to the many instead of the few. Health/adoption and distribute-to-the-many lines are Farzad’s; do not invent the underlying chart/poll beyond what Farzad stated.

Check health QT here. operationdanish here. Distribute post here.

CompleteSkeptic Jev like

CompleteSkeptic, Sep 15 afternoon CT: after co-inventing ChatGPT, spent ~2 years in stealth on RLCD training and a new frontier model releasing today - Jev; claims 20–200x faster, 40–400x cheaper (output tokens free), frontier composable intelligence optimized for decisions; shortest path to AI-based economic revolution (per CompleteSkeptic). Farzad liked. All product/perf claims are CompleteSkeptic’s, not Farzad’s.

Check the like target here.

Starlink V3 - Elon RT + Sawyer Flight 14 stats

Elon Musk, Sep 15 evening CT: Starlink V3 constellation starts being deployed this month; ultimately will deliver more than 100 times the bandwidth of the current 11,000 satellite constellation (per Elon). Farzad RTd. Sawyer Merritt, Sep 16 early morning CT: Starship’s upcoming 14th test flight next week delivers 26 Starlink V3 sats to orbit for the first time and will increase entire Starlink constellation capacity by 3%; would take 10 Falcon 9 launches (277 V2 sats) to add that same capacity; future Starship launches eventually up to 60 next-gen V3 sats adding ~61,000 Gbps in a single launch - roughly as much Starlink capacity as 23 Falcon 9 launches (~640 V2 sats) (per Sawyer). Farzad liked and later RTd Sawyer. Bandwidth/capacity numbers are Elon’s and Sawyer’s respectively.

Check Elon RT here. Elon here. Sawyer like here. Sawyer RT here.

Productivity exhaustion

Farzad, Sep 15 evening CT: AI will likely drive productivity exhaustion in humans because AI systems are tireless and getting better; they will morph to each person’s skill level so non-technical users can push as hard as technical ones; easy output becomes a dopamine hit; great for civilizational productivity at the cost of sanity; weirdly may teach better rest/recovery and fulfillment over hating a corporate job; bullish on humanity’s well-being in the next 5–20 years. Framing is Farzad’s.

Check the post here.