
01
Here’s the Tesla Semi… again
The gist: After nearly a decade of delays, Tesla has begun volume production and first deliveries of its Semi electric truck.
Big pictureThe Semi is aimed at replacing diesel trucks used for long-distance freight, where fuel and maintenance costs matter greatly. Its success will depend not just on its range, but on whether Tesla can produce enough trucks and provide fast charging for busy shipping operations.
Explain like I'm 5Tesla first showed this big electric truck years ago, but it took a long time to make many of them. Now the truck is finally being built in a real factory and sent to customers.
Why it matters nowTesla says production began in April 2026, and first deliveries started this week after an event announcing customers including PepsiCo, DHL, and US Foods. The next test is whether the company can meet announced demand from major freight operators.
Three things to know01The factory targets 50,000 trucks
Tesla says its new Nevada facility is designed to produce 1,000 Semis per week, though the article does not provide current order totals.
02Two range options broaden uses
The Standard Range version is rated for 325 miles when fully loaded, while the Long Range version is rated for 500 miles at an 82,000-pound gross combination weight.
03Efficiency reduced battery demands
Tesla says the production truck uses about 1,700 watt-hours per mile when fully loaded, below its original 2,000-watt-hour target, allowing a smaller battery for the same range.
Trusted source
The Verge ↗

02
There's a new way to break RSA that's faster than anything we've seen before
The gist: Researchers found a faster way to forge signatures against textbook RSA, but the attack does not appear to threaten the padded RSA used by most systems.
Big pictureThe research shows that certain RSA setups are weaker than their key sizes suggest because attackers can use repeated queries to collect clues. Privacy Pass is a notable real-world example, although attacking it would still require an enormous number of token requests.
Explain like I'm 5Think of RSA as a special lock, with padding acting like extra safety gear around the key. The new trick works mainly when that safety gear is missing, so it affects some older or specialized systems rather than most everyday RSA.
Why it matters nowThe newly reported technique reduces the estimated work needed for forgeries, and the researchers say better tools could lower it further. Systems using blind-signature RSA, including Privacy Pass implementations, are the ones to watch as keys are rotated and defenses are reviewed.
Three things to know01An oracle powers the shortcut
The method combines a special number field sieve with an information source called an oracle, using huge numbers of queries to gather enough clues for a forgery.
02The workload is still enormous
For a 1024-bit key, the forgery required about 2^65 operations and 1,380 core-years, compared with an estimated 2^80 operations and 500,000 to 1 million core-years to factor the key directly.
03Privacy Pass faces a volume test
An attack would require roughly 2^43 token requests, while regular key rotation can greatly reduce the chance of success without guaranteeing that it is impossible.
Trusted source
Ars Technica ↗

03
'It almost started a war': US Army nearly boarded a Chinese ship after receiving an "entirely false" AI-hallucinated intelligence report saying it had nuclear arms on board
The gist: An AI-assisted intelligence report falsely identified a Chinese ship’s cargo, bringing armed troops and military aircraft close to a potentially dangerous interception.
Big pictureThe incident shows that AI errors can become especially dangerous when they enter official reports and guide military operations. As the US expands AI use across defense work, people still need to check the evidence before making decisions that could provoke another country.
Explain like I'm 5A computer gave people the wrong answer about what a ship was carrying, and they nearly acted on it. It was like following a faulty map toward a dangerous place before checking the directions.
Why it matters nowThe mistake surfaced during the Iran conflict, after preparations to intercept the vessel were already underway but before the planned operation took place. The episode comes as the Pentagon is rapidly expanding military AI use.
Three things to know01The error entered formal intelligence
The analyst used a chatbot to interpret mixed public and classified information, then used AI again to format the result for circulation among military officials.
02The source material stayed hidden
Officials have not publicly disclosed what cargo the chatbot misidentified, leaving the exact cause of the false conclusion unclear.
03AI checks vary across commands
Different military organizations reportedly use different AI systems and verification practices, so there is no single checking process for every AI-assisted intelligence report.
Trusted source
TechRadar ↗