News
NASA braces for ‘7 minutes of terror’ as rover, rocket crane near Mars
NASA’s most ambitious – and difficult – Mars rover mission to date is nearly at the end of its interplanetary journey, but it is just the beginning of the excitement. On Tuesday (Feb. 16) engineers at NASA’s Jet Propulsion Laboratory (JPL) confirmed that Perseverance is doing well and is prepared to attempt a touchdown at about 12:55 p.m. PST (3:55 p.m. EST) on Thursday (Feb. 18).
NASA’s fifth Mars rover, Perseverance, will attempt a tried and true, but terrifying landing method to reach the Martain surface safely. In a process called entry, descent, and landing (EDL) the rover will burst through the Martian atmosphere at 12,500mph (20,000 kph) and slow to just under 2mph (3kph) in about seven minutes – a process which has earned the nickname “seven minutes of terror.”
From interplanetary spacecraft to Martian rover in seven minutes
For the past seven months, Perseverance has traveled 300 million miles (480 million kilometers) as an interplanetary spacecraft. When it reaches its final destination of Mars on Feb. 18, the spacecraft will have to shed some layers to prepare to land on the Martian surface. Perhaps the most challenging part of the seven minutes of terror is that Perseverance will conduct every aspect autonomously – engineers back on Earth will not be able to intervene due to the communications time delay caused by the distance between Earth and Mars.
During the first stage of landing known as entry, Perseverance will slam into the relatively thin Martian atmosphere at the neck-break speed of 12,500mph (20,000 kph). At approximately 12:38 p.m. PST (3:38 p.m. EST), 10 minutes prior to entering the Martian atmosphere, the Cruise Stage which has reliably propelled Perseverance on its journey from Earth via solar power will separate. This will initiate the official transition from spacecraft into rover.
To protect the rover and its critical hardware Perseverance is housed inside of a protective covering – called an aeroshell – and is outfitted with a robust heat shield. Small thrusters at the crown of the aeroshell help to reorient itself and ensure that the heat shield is facing in the right direction as it enters the atmosphere. The aeroshell and heatshield will absorb and deflect the brunt of the heat energy – reaching about 2,370 degrees Fahrenheit (about 1,300 degrees Celsius) – caused by the friction of entering the Martian atmosphere at such a high velocity.
Once through peak heating and deceleration, Perseverance will utilize a new technology called Range Trigger to determine its exact location and distance to the surface. The spacecraft will utilize this technology to autonomously determine the optimal time to deploy its supersonic parachute – the largest ever sent to Mars – and separate its heat shield. This is expected to occur at 12:52 p.m. PST (3:52 p.m. EST). Once the heat shield has separated the powered descent stage – and the Perseverance rover itself – will be exposed to the Martian environment.
Although a similar descent method has been used in the past with the landing of NASA’s Curiosity rover in 2012, Perseverance’s way of doing things has received a major upgrade.
Once the heat shield has been dispersed, Perseverance will use a radar and cameras to utilize a new landing technology called Terrain-Relative Navigation. Essentially, Perseverance will continuously take images to map out the Martian surface as it descends to determine its exact location. The spacecraft will actively decide and target the best possible safe landing site which can be autonomously changed up to 2,000 feet (600 meters). Then the aeroshell and parachute are jettisoned and it’s the powered descent module’s time to shine.
Using rockets to land, rather than to launch
Just two minutes after ditching the heat shield, at 12:54 p.m. PST (3:54 p.m. EST) and only 1.3 miles (2.1 kilometers) above the surface, the powered descent stage will fire eight throttleable retrorockets to slow the spacecraft’s descent even more and steer it to its chosen landing target. During the powered descent phase, the spacecraft will slow from about 190 mph (306 kph) to just 1.7 mph (2.7 kph).
Once the spacecraft determines that it is 65 feet (20 meters) from the surface by utilizing the Terrain-Relative Navigation, the powered descent stage will initiate the sky crane maneuver. In this phase, the Perseverance rover will be delicately lowered to the Martian surface with a system of Nylon cords.
At 12:55 p.m. PST (3:55 p.m. EST) the $2.4 billion NASA Mars 2020 mission will officially touchdown on the surface of Mars in the Jezero Crater. Once safely down, the sky crane will severe the cords and fly off for a crash landing at a safe distance away from the rover.
During the landing attempt, NASA’s Mars Reconnaissance Orbiter will be overhead and constantly sending telemetry back to Earth via NASA’s Deep Space Network. The telemetry will indicate to engineers back at NASA JPL if the landing procedure was successful and will confirm a touchdown at 12:55 p.m. PST (3:55 p.m. EST).
This will be the first time that a NASA Mars rover will be landing with its eyes open, so to speak. NASA hopes that the first images – and sounds – of the Martian landing will be available to release to the public within about an hour of confirmed touchdown.
Beginning around 11:15 am PST (19:15 UTC) on Thursday, February 18th, NASA will provide live coverage of Perseverance’s landing attempt. The agency will carry the coverage on NASA TV and its website, as well as a number of other platforms including YouTube, Twitter, Facebook, LinkedIn, Twitch, Daily Motion, Theta.TV, and the NASA app. You can view the entry, descent, and landing process in its entirety in the video below provided by NASA’s JPL.
Give the gift of Tesla upgrades.
Shop for someone else or treat yourself. Get more shopping power with a Teslarati gift card.
News
Tesla Full Self-Driving v15: Reality View review, new features explained
Tesla has officially started rolling out the Full Self-Driving version 15 update, and we were lucky enough to get it in the first wave on Saturday night. It comes with a lot of new features, but the biggest is Tesla Reality View, a real-time AI video generation of the outside world that uses less compute than even the earliest versions of FSD driving visualization.
The release notes, which we covered on Saturday, revealed Reality View, Speed Camera Recognition and Slowdown, and a new Bike Rack Mode.
Tesla Full Self-Driving v15.0.1 is downloading!! https://t.co/XRmw9wXaDV
— TESLARATI (@Teslarati) November 21, 2026
After spending plenty of time with the suite over the last 18 hours, I have a pretty good grasp of the new features and what you should expect from them when you get them on your Tesla. CEO Elon Musk said that it will wide release in December with the 2026 Holiday Update.
NOTE: If you’re looking for a review on FSD v15.0.1’s driving performance, I will have a small section breaking down what I noticed over the first drives I took, but I would rather hold off on making any real declarations about it until I have a few hundred miles under my belt. It just feels like I can’t make any real distinctions on anything other than a handful of things Tesla claims in its release notes.
Tesla Reality Vision
Let’s start what I know we are all here for: Tesla Reality Vision is here and it’s absolutely insane. Elon and Co. have also shed a lot of really cool tidbits about the development of it and some of its more intimate details. It’s pretty remarkable what they’re saying about it.
FSD is GTA irl
— Elon Musk (@elonmusk) November 22, 2026
My first impression of Reality View was that it was incredibly detailed. It is honestly hard to believe, at first glance, that there is not a drone following your car. However, once you reach any sort of signage, it’s pretty obvious that it is not a puzzling of camera angles but exactly what Musk says it is:
This is real-time AI video generation (no gaming engine!) using just half of one AI4 TRIP accelerator. Let that sink in …
— Elon Musk (@elonmusk) November 21, 2026
My first rides with it revealed quite a few things.
- It’s very impressive how realistic the texturing of the environment is. The grass, trees, buildings, and all of nature is really well rendered. You can see the shifting as you are moving; I’ve seen a lot of people compare it to the early versions of the Will Smith Spaghetti AI Videos, and that feels like a good comparison. It’s very, very good for a first attempt, and if it looks like this now, it isn’t unrealistic to say that it will look very close to realistic in the future. I would think within the next 8 months is probably somewhat realistic.
- There is some input lag, and Musk has already addressed that by stating: “Currently, the latency is a bit high due to compute constraints, but we have ways to interleave newer UVE features along the diffusion transformer layers to cut roughly 35% of the latency.”
- The lag is not something that is a dealbreaker or even that bad, but it is definitely noticeable. It’s still much better than many OEMs UI, which is as simplistic as menus and has tremendously more lag. It sounds like Musk already has a plan that should alleviate this, so that could come soon.
🚨 Tesla Reality View – First Impressions https://t.co/uSTS94P8Qg
— TESLARATI (@Teslarati) November 22, 2026
One of the coolest features of the Reality View is that it will orient itself based on whether you’re moving forward or backward. When backing out of a space, the drone view switches to the front of the car:
🚨 Tesla Reality View in a parking lot: the “drone” shot moved to the front of the vehicle when in reverse, and moves to its regular position upon FSD moving into Drive https://t.co/lbPGkfa1Oq
— TESLARATI (@Teslarati) November 22, 2026
Another really nice touch is the quieter turn signal chimes, which start at a normal volume and then subtly get quieter until they are barely noticeable. It’s a very peaceful touch that Tesla added with the v15 update:
🚨 Tesla told the Turn Signals 🤫 🤐 on FSD v15 https://t.co/I3lE8s7Ei9
— TESLARATI (@Teslarati) November 22, 2026
First Thoughts on Driving Performance
Overall, I have only noticed a handful of things. Driving performance feels very similar to recent FSD versions, but these are two things I caught:
- When taking a left-hand turn from the middle of an intersection with a lot of room, the car didn’t just turn left, it actually turned right a little bit and then turned left. This is something you see a lot of very inexperienced drivers do; they’ll be in a Honda Civic and they act like they need to swing it wide to make a very straightforward turn. I have never seen FSD do this. This turn is one I take each day, as it is on my way to the gym. It was absolutely crazy and I was really thrown off by it. I hope this isn’t a new thing.
- I didn’t get to see many potholes but I did see a lot of manhole covers, and FSD didn’t do anything to avoid them. In our YouTube video down below, you’ll notice it runs over many manhole covers. I was hoping it was try to avoid these, but I didn’t seem to get that impression.
In the coming days, I’ll do a longer review on the driving performance and whether or not v15.0.1 includes the things that Tesla explicitly mentioned in the Release Notes.
Elon Musk
Tesla Full Self-Driving v15 rollout begins: full release notes and improvements
Tesla has announced that Full Self-Driving version 15 is now rolling out, after CEO Elon Musk said earlier today that the wide rollout would occur in December.
Now, Tesla is rolling out to Early Access Program (EAP) members v15’s first builds, and there are a lot of new improvements.
Tesla explicitly stated that safety improvements come in the form of “25x better handling of safety critical cases.” It is also rolling out “Tesla Reality View,” a new version of the Driver Visualization that appears to be much more realistic and could potential use Unreal Engine, according to the teaser video Tesla posted on social media. This “adds a live visualization of the drive that lets you see & control what FSD is planning to do.”
FSD v15 now starting to roll out – upgrading every part of the end-to-end driving network
Safety
25x better handling of safety critical casesTesla Reality View
Adds a live visualization of the drive that lets you see & control what FSD is planning to doNavigation for start & end of the trip now learns end-to-end from fleet route preferences. Later v15 releases will expand this to the entire route.
You also should notice better detection of signs, animals, debris, potholes and standing water, plus more natural speed handling in school zones and inclement weather.
Lastly, FSD can now slow for speed cameras & Bike Rack Mode lets you enable FSD w/ bike racks or similar hitch accessories
— Tesla (@Tesla) November 21, 2026
Perhaps the biggest improvements that Tesla mentioned, however, are that of Navigation fixes and better detection and aversion from potholes, debris, and animals, as well as “more natural speed handling in school zones and inclement weather.”
Tesla also added a new “Bike Rack Mode,” which will enable FSD’s use when bike racks or other hitch accessories are attached.
Tesla also revealed the full Release Notes for v15:
- Upgraded vision encoder shared between FSD, Robotaxi, and Optimus, which is trained on millions of internet videos for improved generalization and data efficiency, improving the critical intervention rate by 25x.
- Introduced Tesla Augmented Reality, replacing all of the explicit model signals into a single end-to-end visualization using less total compute.
- Improved navigation by using an end-to-end learned model based on route preference data from the fleet.
- Improved parking location selection by using personalized parking history, and the map will now display multiple parking options at your destination.
- Improved driver monitoring by increasing vision-based attention monitoring availability with hats and sunglasses and decreasing strictness in higher speed driver profiles.
- Improved detection of road signs, small animals, road debris, and chains.
- Improved detection of potholes, road details, and avoidance of deep standing water.
- Improved speed handling for each driver profile, school zones and speed limits, and inclement weather.
- Improved Autoshift functionality by unifying with FSD-based gear predictions.
- Tesla Reality View New: real-time generative driving visualization that continuously streams a view from an arbitrary virtual camera situated above your vehicle, enabling you to see and control what FSD is planning to do. Select between fullscreen visualizations, maps or a split view – or choose quiet mode for a calming cabin experience. Upcoming navigation turns display directly in the driving visualization. Turn signals are quieter and fade out after a few clicks.
- Speed Camera Slowdown: FSD (Supervised) can now slow down when it detects a speed camera. Press the accelerator or increase your speed profile to cancel the slowdown. If you cancel twice at the same camera, your Tesla saves that location and will not slow down there again. To remove saved cameras, go to Controls > Service > Manage Saved Locations.
- Bike Rack Mode With Bike Rack Mode, you can now engage FSD (Supervised) when a bike rack or similar hitch accessory is installed. When Bike Rack Mode is enabled, FSD (Supervised) pulls over near your destination and does not reverse. Autopark, Summon, and Backup chimes and warnings will be disabled.
It sounds like there are some really great improvements, and this is probably just a handful of what other improvements will come from these upgrades. For example, the UI changes will likely come with some other features that will be discovered upon the public having hands-on experience with it.
Musk said this morning that a wide release would occur sometime in December.
We’re looking forward to testing it when it arrives to us!
Elon Musk
Tesla Full Self-Driving v15 gets release date from Elon Musk
Tesla Full Self-Driving v15 has received an official release date from CEO Elon Musk, who says the newest and most capable version of the suite will go wide in December with the 2026 Holiday Update.
That means the first iterations of the suite could actually come sooner, as the Early Access Program is the initial batch of owners to receive a new FSD release. Musk’s comments released a few tidbits of information, but nothing overwhelmingly revealing about its capabilities.
Musk was talking about supervised fine-tuning, stating that it is like “getting a haircut with a chainsaw,” before stating that FSD v15 will roll out wide for owners next month:
SFT is more like getting a haircut with a chainsaw than it is “fine-tuning”! Should just be built into primary training.
And having a “system prompt” is silly! Default behavior should be sensible.
The above are already true for Tesla AI (not yet deserving of “super”). It is the most efficient AI on Earth.
FSD V15 goes to wide release next month. It will be something special.
— Elon Musk (@elonmusk) November 21, 2026
Musk’s Confirmed Changes and Fixes
Musk simply confirmed the new Full Self-Driving version would roll out with the 2026 Holiday Update, and the only two explicit changes he confirmed were updates to the User Interface (UI) and two small changes.
Musk’s comments on the UI confirm there are big changes coming to it, as he stated:
@roger_kappler Yeah.
The new user interface will re-render reality in a mind-bending way. Hard to describe, but it’s awesome.
— Elon Musk (@elonmusk) November 21, 2026
The UI changes will likely upgrade AMD Ryzen-equipped Tesla vehicles running AI4 to visualize vehicles using the Unreal Engine, something that was found in coding last year:
Tesla plans to use Unreal Engine for driver visualization with crazy upgrade
Additionally, one user stated that their biggest complaints with v14’s most recent releases is that they have to “stop refusing to use the HOV lane” and last-second lane changes to merge or turn happen too frequently. Musk said these should be fixed in v15.
@riley_stahura Should be fixed in V15
— Elon Musk (@elonmusk) November 21, 2026
In the past, we’ve also gotten small bits of detail about the v15 release, including from Tesla’s Head of AI, Ashok Elluswamy, who stated back in September that there are improved safety and collision avoidance systems coming with the v15 release.
Tesla plans big safety improvements for Full Self-Driving v15
The big prediction for v15 is whether this will enable at least some unsupervised driving in personal Teslas. There are likely to be some massive jumps in capability, although FSD has already proven itself to be a great way to travel. Nevertheless, there are a lot of things it could do better.
FSD still struggles with parking, it has routine bugs from update to update, it does not recognize potholes or large bumps on the road, speed control is a common complaint, and Navigation has been one of the biggest pain points.
However, v15 will likely bring massive improvements to capability and give people more options, like potentially parking orientation and more refined spot selection. We can only wait and see what Tesla will include.
