Destination voice is not a work session

The new framing is easy to find: owners saying 'Hey Grok, take me to the ATM' or 'take me home' under FSD Supervised, and Tesla amplifying those clips as what it feels like to have your own robotaxi. On Cybercab, cabin Grok handles climate, media, and destination tweaks by voice. Both are real. Neither turns a supervised drive into Riding mode. If there is a steering wheel and you are responsible for it, you are still the legal driver — full stop. But destination voice is genuinely useful in Driving mode, and it is worth being precise about what it buys you.

What destination voice actually buys you

• Navigation load off your working memory. You stop juggling route changes, reroutes, and the touchscreen mid-drive. That frees attention for the road — which is the entire point.
• A clean window for voice capture. Once the car owns the route, a hands-free stretch becomes a good slot to think out loud, dictate notes, or rehearse a difficult conversation — all legal and glance-safe.
• Better arrival prep. Ask for the destination early and you get the last few minutes to close the loop: dictate the two-line summary, set the follow-up you will send from your desk, park with a plan instead of a thread you abandoned mid-sentence.

Notice what is on that list: capture, thinking, and prep. Nothing that requires your eyes off the road or your hands off the wheel. That is the honest ceiling for Driving mode, and it is higher than most people use.

The three-phase ritual

Voice-only, all of it, because you are the driver.

• Before you roll: set the destination by voice, then dictate the one outcome for this drive — the decision to untangle, the draft to think through, the call to rehearse. Thirty seconds on the curb beats a scattered half hour in motion.
• While rolling: think out loud. Use the assistant as a thinking partner, capture fragments as they come, and let the car handle routing. If traffic forces a change, you say 'take me home instead' and keep going — no touchscreen, no detour into your phone.
• On arrival: use the last two minutes to wrap, not to start. Dictate the summary, the to-do, the single next action. Save nothing for 'when I get to my desk that I will have forgotten by then.'

The red lines do not move

• No reading threads, email, or documents while the car moves — supervised autonomy still means supervision.
• No typing, on any screen, for any reason.
• No video calls and no watching anything.
• The center display is a control surface, not a work surface. If you are touching it to read or respond to work, you have crossed the line regardless of what the car can steer around.

The 'your own robotaxi' framing invites exactly this slippage: if it feels like a chauffeur, a laptop starts to feel reasonable. It is not. A distracted driver with good driver assistance is still a distracted driver, and the assist disengages at the worst possible moment.

For the contrast: a real robotaxi or Cybercab leg — where nobody in the cabin is the legal driver — is Riding mode, and that is where the laptop, the calls, and the deep-work blocks live. We keep the split sharp because blurring it is how someone gets hurt; the full breakdown is in Supervised FSD is not Riding mode.

Destination voice takes the driving off your mind. It does not take you out of the drive.

Use the errand pattern — it is one of the best things in the car today. Just keep it voice-only, keep the ritual above, and remember why people trust this site: we do not sell a mode that the law has not granted you yet. Trust is the moat.