Here’s one metric for tracking SpaceX’s IPO later this week: The company has changed the venture industry’s perspective on long-term, capital-intensive space so much that a talented founder with no space experience can fund a space data center company.

Orbital, a new firm that emerged in May from a16z’s startup accelerator program Speedrun with a $5 million seed round, is the latest company promising to do inference in space — just as soon as Starship is flying regularly. Other investors include Basis Set, Human Element, Wayfinder, Antler, Anti Fund, Ascent, Rubik, Zero Knowledge Ventures, LYVC, Feld Ventures, New Legacy, FNDR, UpHonest and Asterisk.

Founder and CEO Euwyn Poon previously founded e-scooter company Spin in 2017 and sold it to Ford a year later, joining the automotive giant. When he was ready to start a new company, a16z’s Speedrun was eager to get on board, according to partner Andrew Chen, who told TechCrunch that Poon worked through several ideas before landing on space data centers.

You’re familiar with the pitch. There’s insatiable demand for AI compute, and deploying it is slow going on Earth. Why not head to space for limitless sunshine and limited environmental reviews? The main problem is the brutal economics of launching stuff into orbit, which currently leaves the business case unable to close.

Orbital, like many of it competitors, is betting on SpaceX figuring out its Starship rocket and offering it to commercial customers. “We will get to full scale when Starship comes online,” Poon explained. The price of the Falcon 9, the current state of the art, “makes this not economically feasible.”

For now, Poon and company — which includes about a dozen folks in Los Angeles, with experience at Amazon LEO, SpaceX, and Northrop Grumman — are working toward a demo flight that will see the company fly an Nvidia Blackwell chip on a partner’s satellite to test Orbital’s radiation shielding and thermal management tech. In 2028, the company hopes to launch its first data-processing spacecraft with Nvidia’s Space-1 Vera Rubin-class GPUs.

At that point, the company wants to start doing piece-wise inference work, which would allow it to generate revenue with each satellite launched. That’s a similar path to rival data center start-up Starcloud, which already has a GPU in orbit and plans to launch several more to generate income until Starship enables them to deploy their full constellation.

Orbital’s goal is to deploy 10,000 satellites that provide a distributed gigawatt of computing power, with each satellite providing 100 kw of power. For comparison, Elon Musk said SpaceX expects its AI satellites produce up to 150 kw, and Starcloud expects to field larger 200 kw-rated spacecraft to run chips.

Some companies are too impatient to wait for Starship. Cowboy Space Company, another space data center startup backed by a16z, recently decided to start building its own rockets. Jeff Bezos’ space company Blue Origin also announced plans to launch data centers into space using its New Glenn launch vehicle.

Poon is confident that the breadth of AI demand will allow many companies to succeed. “There’s so many lanes for companies in our space to pursue,” he told TechCrunch, before rattling off an array of choices that included companies pursuing different AI workloads, designs, and concepts of what an space data center looks like.

Chen said that Poon’s experience scaling up a company that deployed 250,000 scooters across 100 cities shows he can manage the tricky task of building an aerospace company. Over the long term, a project like this might take a decade and $5 billion or more, but Chen said venture firms are more comfortable with timelines like that.

“This kind of thing would have sounded crazy 10 years ago when we were all building mobile apps,” he said. “Starting it in 2026 just lets you tap into all the energy and excitement that’s that’s happening in the capital markets.”

Poon found his way into the space data center business by a circuitous route. After leaving Ford, he bought a Nvidia A100 on a lark, co-locating it in a Santa Clara data center and serving open-weight models. That first-hand experience convinced him the value in delivering compute in the era of AI.

Now he’s just got to put a couple thousand GPUs in space.

Photo credit: Delft University of Technology | Micro Aerial Vehicles Lab
Researchers at Delft University of Technology have built a navigation method for small drones that copies one of nature’s most efficient routines. Honeybees fly long distances along twisting routes yet still head home with striking accuracy. The new system lets a drone do something similar after just one short practice flight near its base, all while using a learning program small enough to fit in the memory of a basic phone app icon.

Honeybees keep track of distance and direction by observing how the ground and surroundings move past their eyes. Then, over time, slight inconsistencies begin to appear. To remedy this, bees will frequently undertake a quick small learning flight shortly after starting feeding, looping around the area near the hive and gathering all of their visual memory of what is around them. Later, when they’re far out in the wild, their memories guide them back to the hive, even though they can’t see it.

Sale

DJI Neo, Mini Drone with 4K UHD Camera for Adults, 135g Self Flying Drone that Follows You, Palm Takeoff…

• Due to platform compatibility issue, the DJI Fly app has been removed from Google Play. DJI Neo must be activated in the DJI Fly App, to ensure a…
• Lightweight and Regulation Friendly – At just 135g, this drone with camera for adults 4K may be even lighter than your phone and does not require FAA…
• Palm Takeoff & Landing, Go Controller-Free [1] – Neo takes off from your hand with just a push of a button. The safe and easy operation of this drone…

A group of Delft researchers wanted to see if a drone might follow a similar path. Rather than develop an incredibly thorough three-dimensional picture of the entire area, as if a robot were trying to recall a complex plan, they just had the thing snap a few short panoramic photos during a quick learning fly around the starting position. They then uploaded the photographs to a small learning system, along with the drone’s initial estimates of how far and in which direction it had traveled. Yes, those calculations become a little jumbled, but the program still manages to extract enough information from the surroundings to determine its home territory and return the drone to that location.

When the drone takes off on a longer flight, it primarily relies on those motion estimates to get there and back, right up until it approaches the place where it learned to fly. The visual portion of the system then takes over and returns the drone to its original location. It also teaches the drone to fly faster when it is still in the wilderness and slow down as it approaches, similar to how bees adjust their own speed near the end of a journey.

Tests in these huge indoor environments were consistently successful. The drone might take off from a certain position, fly all the way to the other side of this massive hangar or test arena, and then return without ever seeing its home base. Outside, the same method worked fine for a drone that flew more than 600 meters before turning and returning home, and it continued to work well even on calm days, but it began to get a little wonky on windy days, primarily because gusts would blow the drone around and interfere with the camera’s view.

The outdoor version works on a learning system that uses only 42 kilobytes of memory, reducing memory utilization to a minimum. Traditional map-building methods often require substantially more memory and processing capacity, pushing designers to invest more in drones that use more energy and are more expensive to fly. The bee-inspired technique keeps the technology lightweight, allowing small drones to safely fly around people or into places like greenhouses where they don’t want to clash with anything.

Crop monitoring has a clear advantage because a lightweight drone can fly between rows of plants looking for early signs of disease or pests without the bulk or risk associated with larger mapping systems. The same navigation technique makes it perfect for robots moving about warehouses inspecting industrial equipment or functioning in places where satellite signals are distorted due to trees, buildings, or weather.
[Source]

DINUM, the digital affairs directorate of the French government, warned that hackers used a hijacked user account to breach Tchap, the French government’s encrypted messaging platform.

Developed in-house by DINUM in collaboration with ANSSI (the French Cybersecurity Agency) in 2018, Tchap is an instant messaging service and collaboration tool based on the decentralized Matrix protocol, designed exclusively for the French public sector.

Tchap has now reached over 300,000 monthly users and over 500,000 downloads on Google’s Play Store after Prime Minister François Bayrou mandated the use of Tchap and banned foreign apps for work communications for all civil servants in early August 2025.

DINUM revealed on Monday that ANSSI detected a Tchap breach on Sunday and said that a threat actor gained access to the secure instant messaging platform using a compromised user account.

The French digital affairs directorate has also alerted France’s data protection authority, the CNIL, to the incident due to the potential exposure of personal data shared by some users in conversations that the attacker could access, and has alerted all Tchap users, reminding them that public chat rooms are accessible to any user and are not encrypted.

“At this stage, the account originating the malicious requests has been identified. It was immediately blocked to remove the attacker’s persistent access and allow for a thorough analysis of the data they were able to access. The investigation continues, including the study of event logs, to identify the conversations that the attacker was able to access and the nature of the exfiltrated data,” DINUM said in a Monday press release.

“A message has been sent to all Tchap users reminding them that a public chat room can be found and joined by any user and that its content is not encrypted. In accordance with Tchap’s terms of service, no personal, sensitive, or confidential information should be exchanged in public chat rooms: such exchanges should be reserved for private chat rooms.”

While the DINUM has not shared any further details regarding this breach, a threat actor claimed responsibility for the incident over the weekend, shared a sample of stolen files, and said they gained access to the platform following a social engineering attack.

​”I social engineered a valid account on the education shard (matrix.agent.education.tchap.gouv.fr). Everything below is what that one account could reach, other shards will have more,” they said.

​They claim to have stolen hardcoded LDAP credentials allegedly leaked via a PowerShell script shared by a French tax authority regional director and over 13.5GB of documents and media files shared by public servants using the Tchap service.

The threat actors also allegedly scraped nearly 650,000 messages and information on over 73,000 accounts, including email addresses, organization information, meeting links, and account and device metadata.

“Every file ever shared on Tchap, on any shard, is downloadable without a token,” they added. “The media IDs come from the messages. Once you have a message with a media URL you can pull the file freely regardless of which shard hosts it.”

BleepingComputer reached out to DINUM with questions about the incident, but a response was not immediately available.

Last month, French authorities detained a 15-year-old suspected of selling data stolen in an April cyberattack on ANTS (Agence nationale des titres sécurisés), the country’s agency for issuing and managing official identity and registration documents.

Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.

The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.

Get the whitepaper

One unique feature is a free “moving artwork” functionality. Imagine a static painting where a portion of the art, such as a river or mist over a mountain, is moving slightly. It’s calming and beautiful. At least dozens are available, but I’m not able to add my own. The TCL NXTVISION also features paintings in a slowly moving video format, but the video art on the Ember Artline looks more realistic. In particular, I was blown away by some of the subtle clips, including one scene that shows the sun passing over a mountain range in extreme hi-def.

Another unique feature is Match the Room. It uses a photo where the TV is situated, then generates AI images that fit within the space’s color scheme. Having my TV display art that fits the aesthetic of my family room’s brown walls and dark blue couch made me feel like an interior designer.

Then there’s the Omnisense feature, which claims to know when you leave the room. At that point, it turns off the art display, then powering it back on when you return. (For the sake of comparison, the TCL doesn’t have a motion sensor; most Ember models and the Frame Pro do have motion sensing.) During my testing, the Omnisense worked about 90 percent of the time, sometimes failing to clock my presence in the room and not turning on the Artline as it should. Amazon reps told me they have not heard this feedback before. Even so, the feature was mostly functional and is a great power-saving capability.

The Alexa+ voice search function proved limited on the Artline, though you can conveniently use it to raise the volume and start apps by voice. Unfortunately, I wasn’t able to use it to search for classic masterpieces, like Van Goghs or Rembrandts. However, the virtual assistant could answer questions about the weather, play music, provide sports scores, and even chat with you about politics. I loved using Alexa+ to find movies and shows. I made complex requests like “show me every thriller from the last two years that has an 80 percent score or higher on Rotten Tomatoes,” and that actually worked. You can also show Amazon Photos, animated art, and use Alexa+ to generate AI artwork. The AI artwork is a bit generic. That said, using voice commands to make art is unique and fun.

The main drawback of the Ember Artline compared to the Samsung Frame Pro is that the contrast is a bit washed out. I couldn’t find Vincent Van Gogh’s The Starry Night when I scanned through the Artline’s available artwork, so I loaded it as a photo. When I did this on the Frame Pro, the incredible contrast and color quality provided a vibrant, realistic texture, like you could reach out and feel the grooves of thick brushstrokes. On the Artline, the same painting lacked texture and contrast.

Screen Play

Photograph: John Brandon

I was surprised to discover that the Ember Artline doesn’t use new screen technology—it’s roughly the same QLED tech as previous Amazon Fire and newer Ember televisions. Which is to say, the contrast ratio (which Amazon does not release) is not outstanding, based on my experience viewing multiple movies and shows. There’s a new anti-glare matte finish, though, that makes all of the artwork (and whatever content you are streaming) look more realistic.

• Broadcom shares plummet after it reports earnings beat for Q2 2026
• The company’s market cap shrank a mammoth 19% across two sessions after its earnings report was scrutinized by Wall Street, with its AI chip sales outlook being notably softer at $16 billion versus $17.2 billion for Q3
• The behavior may indicate a frothy AI market, where investors continue to obsess over future guidance and valuations versus current performance

Broadcom’s Q2 2026 earnings report, by all means, reads like that of a company that is gearing up to benefit from a splurge in AI spending across the next decade.

It posted record revenue, tripling its AI chip business year-on-year and topping its Q2 earnings-per-share (EPS) guidance.

from the doge-screwworms-america dept

Back in March of 2025, when Elon Musk was spearheading his bullshit DOGE non-agency and running around cutting funding to all kinds of government programs under the notion that they were a blatant waste of taxpayer money, the cuts were so obviously haphazard and ill-informed that it was making everyone’s head spin. Then, after cuts to funding and staff were done, the very same Trump administration began scrambling to restore it after government agencies found themselves unable to function properly as a result.

But not every bit of funding was restored. Part of what DOGE cut was roughly 5,000 funding grants for USAID. One of those grants funded a detection and warning program for, you guessed it, screwworms!

Following the U.S. Department of Agriculture’s confirmation on Wednesday that the New World screwworm fly has reached south Texas for the first time in decades, questions are being raised about what role DOGE cuts played in what could become a crisis to the nation’s cattle industry.

The Elon Musk-led Department of Government Efficiency gutted the United States Agency for International Development, which included a program dedicated to preventing the spread of the parasite across the U.S.-Mexico border, according to a report from Agri-Pulse published last March, which cited a list of cut programs sent to Congress.

The screwworm prevention program was part of roughly 5,300 grants and programs cut from USAID. The program also monitored outbreaks of avian flu in Asia, according to the report.

Advertisement

Everyone who was paying attention knew then that screwworm flies were a major problem. These parasites infect livestock, pets, and sometimes humans by burrowing into their flesh to lay their eggs, potentially killing the host in a matter of weeks. A mass-infection event would decimate the livestock of cattle specifically in America, which is already at an extremely low level. That reduction in cattle counts occurred because there are fewer ranchers today than before, which is itself due to the increasing costs of raising cattle in America. Those costs would be for fuel, parts for ranching equipment, and fertilizer. And you can explain most of those rising costs on a combination of tariffs and our idiotic maybe-war with Iran.

Meanwhile, the cost of beef has been further driven higher because of a USDA import ban on Mexican cattle that has been in place since May of 2025. You’ll never guess why that ban was put in place.

“The United States has ordered the suspension of livestock imports through ports of entry along our southern border after the continued spread of the New World Screwworm in Mexico. Secretary Berdegué and I have worked closely on the NWS response; however, it is my duty to take all steps within my control to protect the livestock industry in the United States from this devastating pest,” said Secretary Rollins. “The protection of our animals and safety of our nation’s food supply is a national security issue of the utmost importance. Once we see increased surveillance and eradication efforts, and the positive results of those actions, we remain committed to opening the border for livestock trade. This is not about politics or punishment of Mexico, rather it is about food and animal safety.”

But those increased efforts never came to pass, in part because DOGE cut the funding for them. This government doesn’t have the ability to claim they didn’t know screwworm flies were a problem. Its own USDA said it was. The Trump administration can’t claim it isn’t responsible for the reappearance of the parasite or blame the rising cost of beef and milk on someone else. Trump’s tariffs and war with Iran are directly responsible for it, and any significant issue with screwworm flies will cause that cost to rise even further.

Hell, ranchers have apparently been screaming about this to try to get the government’s attention for months and months now.

Agriculture officials and cattle industry leaders raised alarm about the cuts at the time and, for the last several months, pleaded with the government to step in as they monitored screwworm infections moving north through Mexico—but they were ignored, Texas Agriculture Commissioner Sid Miller told NBC News.

And now we’re here, once again, with another once-eliminated issue that will plague the American people due to this administration’s rank incompetence. Just like the measles. This government is leading us backwards.

So what’s the plan for screwworm flies now? Fly-sniffing dogs and the release of millions of sterile male flies to crowd out the parasite’s ability to reproduce. And that actually is the proper plan to combat this thing… long term. But not in the immediate, which is why we’re likely to see it spread.

The plan to prevent a US outbreak of the New World Screwworm focuses on deploying hundreds of millions of genetically-altered sterile flies. Experts, though, say the supply of sterile flies is too low to immediately impact and halt the growing screwworm population.

60 years our country has been without the New World Screwworm. And now it’s back. Because this government would rather do government cut dinner theater than the hard work of governing.

So they next time you’re at the grocery store and can’t believe the price of a steak, you can thank the Trump administration for it.

Filed Under: cattle, doge, elon musk, screwworm, trump admininstration, usaid, usda

iOS 27 has the usual Apple headline grabbers, led by Siri AI, Apple Intelligence upgrades, and broader system improvements coming this fall. But some of the best day-to-day changes sit much lower on the feature list.

Apple is making smaller changes across Mail, Photos, Home, accessibility, connectivity, AirDrop, and even alarms. They won’t dominate keynote clips, but they could remove the tiny annoyances people run into long after the novelty of a major update fades.

Which fixes sound instantly helpful

Start with Mail, because search is one of those features people only notice when it fails. iOS 27 adds a new ranking system designed to push more relevant results to the top, which should help when a receipt, boarding pass, or appointment confirmation is buried under years of inbox debris.

The other sleeper fix is smoother network switching. iPhone will more seamlessly choose between Wi-Fi and cellular, which could help when you’re leaving home with Maps running or trying to keep a FaceTime call alive while moving between networks. AirDrop and AirPlay are getting speed boosts too, with faster transfers and quicker nearby-device discovery. Bad handoffs are tiny failures, but they make an expensive phone feel dumb fast.

What changes for shared photos

iCloud Shared Albums are getting friendlier for mixed-device groups. Android and Windows users will be able to join and contribute more easily through iCloud.com, which should help families and group chats that don’t live entirely inside Apple’s ecosystem.

Apple is also adding full-resolution sharing, filtering, reactions, more invite options, and expiring Shared Albums. That’s overdue housekeeping for a feature that feels simple until a trip ends and everyone starts asking where the photos went. The Photos app is getting another handy trick too, with an option to save a still image from a video frame.

Which upgrades are easy to miss

Accessibility gets some of the strongest smaller additions in iOS 27. VoiceOver can provide richer image descriptions, while a new captioning feature can generate synchronized subtitles for videos. It can also translate existing captions, which could make saved clips and shared videos easier to follow without a separate app.

The tiny utility pile keeps going. Alarms, timers, alerts, and system sounds can be separated from ringtone volume, so one setting doesn’t have to govern every noise your iPhone makes. Notes can copy and paste Markdown, which is deeply unsexy unless you write for the internet, in which case it’s basically a tiny holiday.

These aren’t the loudest iOS 27 changes, but they’re the kind that may make the phone feel a little less needy every day.

Families in Britain approached a major household decision with a mix of excitement and uncertainty in the spring of 1998. A BBC program from the Computers Don’t Bite campaign followed two of them step by step as they bought their very first home computers. The episode captured the practical choices, the new vocabulary, and the hands-on setup that turned an expensive piece of equipment into something the whole household could use.

The Lyons family went to a specialty electronics store in search of a system that could handle the kids’ school projects, keep track of the household finances, and make some simple graphs without making them nuts. Jenny Lyons spoke with the salesperson about their needs, and they were guided to a Compaq 4620. This one contains a fast 266 MHz Pentium II processor, 48 MB of RAM, a 4 GB hard drive, and a built-in modem to accommodate the growing number of individuals who want to go online. Windows 95 was pre-installed, and the full bundle, including a printer, cost less than 1000 pounds. The Lyons bought a beautiful, ready-to-use system that promised plug-and-play capabilities right out of the box.

Sale

GEEKOM Air12 Mini PC for Business & WFH (2026), Intel 7505 (Up to 3.5GHz), 8GB RAM (Dual-Slot…

• [Unmatched Productivity for Office & Dev Workloads]: Powered by the Intel Pentium 7505 (25% faster single-core performance than the N95/N100/N…
• [24/7 Silent Stability & Metal Durability]: Built for non-stop performance. The Air12 features a strategic metal mid-frame, a 3x-reinforced shell, and…
• [Triple 4K & 8K Workspace Hub]: Transform your desk into a professional command center. With support for up to three 4K displays or one stunning 8K…

The Plummer family took a different approach, first looking through computer periodicals before purchasing a machine directly from the vendor. This was the material they felt most at ease with. They ended up with a system with similar fundamental specifications: a 200 MHz processor, 32 MB of RAM, a 4 GB hard drive, and a color printer, all for less than 1000 pounds. The Lyons and Plummers both ended up with working home systems, but their choices were influenced by their own comfort levels.

Sue Davis and Dave Green tried to simplify the terminology in the sales brochures. RAM is essentially the temporary workspace that the computer uses while turned on; the CPU runs the apps from there. The ROM stores the machine’s permanent instructions for starting itself up. The hard drive holds saved data, letters, games, and photos. A family machine need adequate room to prevent games and paperwork from filling it up too quickly. Clock speed, defined in megahertz, is important since it determines how quickly the CPU can accomplish tasks, with numbers in the 200–266 MHz suggesting that the CPU will be much quicker than previous devices.

By the time the shoppers left the store, they had a decent idea of what to look out for. More megahertz is better, implying faster labor. Ordinary tasks will run smoothly with at least 32 MB of RAM, and 48 MB is a reasonable amount of space to spare. Hard drives with a capacity of 2 GB will sufficient for most basic operations, while 4 GB will provide some additional room.The presenters also underlined the importance of sticking to a reasonable budget, since £1,500 is about the maximum unless someone wants to get into professional creative work. They also recommended speaking with knowledgeable friends or reading a few periodicals before making a decision.

Setup day arrived for the Plummers, and what a sight it was: a box full of boxes containing everything they had bought. There was a monitor, the main computer unit, the printer, some speakers, a keyboard, a mouse, and a few cords. The color-coded cables made things a bit easier. They followed the steps in a manufacturer video and were quickly up and running on Windows 95. A quick test with Solitaire revealed a tiny glitch that was readily remedied by resuming with the trusted Ctrl-Alt-Delete combination or turning off and back on. After a brief delay, a call to the helpdesk supplied them with fantastic service.

Then followed software installation, with Microsoft Works coming on a CD-ROM and booting up with on-screen instructions. The application offered word processing, spreadsheets, and a basic database. The templates were all ready for use. One presentation showed a family letterhead all filled out with names and addresses, ready to be printed. Ordinary jobs become considerably more manageable without having to choose a program to conduct each activity separately.

offbeat

Prochain arrê: Gare du Bork! French capital city train does the tech can-can

BORK!BORK!BORK! Good news from the Paris Metro. To show off the nation’s technological prowess, Parisian techies have eschewed such fripperies as advertisements and transit information in favor of a good, old-fashioned directory browser.

Spotted by eagle-eyed Register reader “DJ Finsletown,” the screen adorned a carriage of Line 11 at the Chatêlet terminus earlier in June. It looks to us as though a web server threw a bit of a wobbly and showed users its undercarriage.

Then again, France did give us the can-can dance, in which participants perform high kicks that reveal their underwear, so perhaps this is simply the tech version of the music hall standard.

Our reader pondered whether the system behind the scenes was struggling with a recent heatwave, or the borkage was the result of some over-exuberance by football fans following PSG’s recent Champion’s League final victory.

What is perhaps more likely is that some techie in a backroom somewhere opted for a Gallic shrug rather than reaching for the inevitable CTRL-ALT-DEL to revive the stricken system.

Line 11 is notable for being one of the least-used lines of the Paris metro system until a recent extension. It also featured some of the oldest-running stock and was notable for rubber-tired trains and heat that might make London’s Central Line commuters mutter “steady on.”

It was also one of the last lines to be constructed in central Paris, on the former route of the Belleville funicular. The funicular, which ran along the streets in a manner familiar to San Francisco residents, ceased operation just over 100 years ago, in 1924. Line 11 got the nod in 1935.

As for the web server disgorging its contents, that is a far more modern innovation. Heck, it might even be some AI advert demonstrating what might, or might not, happen if an agent is let loose on a system. In the case of the latter, a borked display would, we suspect, be the least of the problems faced by passengers.

Or, as our reader joked, “Probably likely the train’s IT systems just went on strike, like all good French fonctionnaires!” ®