Lightning-Proof Tech and OpenAI's Pentagon Partnership: The Startups and Deals Reshaping the Industry
This is today's edition of The Download, our weekday newsletter delivering a daily briefing on the most consequential developments in technology.
This startup claims it can neutralize lightning and prevent catastrophic wildfires
Startup Skyward Wildfire is making a bold claim: that it can prevent large-scale wildfire events by intercepting the lightning strikes that trigger them. The company has yet to publicly disclose its methodology, but internal documents indicate it is building on a cloud-seeding technique the US government first explored in the early 1960s — dispersing metallic chaff, specifically narrow fiberglass filaments coated in aluminum, into storm systems to disrupt electrical discharge.
The company recently secured a multimillion-dollar funding round to accelerate development and scale its operations. Yet the scientific community remains cautious. Researchers and environmental analysts point to a range of unresolved questions: how reliably the seeding performs across different atmospheric conditions, what volume of material would need to be deployed, how frequently interventions would be required, and what downstream ecological consequences might emerge from repeated use. Read the full story.
—James Temple
OpenAI's Pentagon "compromise" is precisely what Anthropic refused to accept
OpenAI has formalized an agreement granting the US military access to its technologies within classified operational environments. CEO Sam Altman acknowledged that the negotiations — initiated only after the Pentagon publicly rebuked Anthropic for declining similar terms — were "definitely rushed."
The company has been careful to frame the deal as a principled compromise rather than a capitulation. OpenAI published a blog post asserting that the agreement includes explicit safeguards against deployment in autonomous weapons systems and large-scale domestic surveillance operations. Altman maintained that OpenAI did not simply accept the same contractual terms Anthropic had rejected.
Whether those assurances hold in practice is another matter. As the military accelerates a politically charged AI strategy amid active strikes on Iran, it remains unclear whether OpenAI can meaningfully enforce the safety guardrails it has promised — or whether that will be sufficient for employees who pushed for a firmer stance. Navigating that tension will require considerable dexterity. Read the full story.
—James O'Donnell
This story is from The Algorithm, our weekly AI-focused newsletter. To receive stories like this before anyone else, sign up here.
The must-reads
A curated selection of today's most significant, surprising, and thought-provoking stories from across the technology landscape.
1 Gulf states are in a race against time to intercept Iran's drone offensive
Interceptor stockpiles could be exhausted sooner than anticipated. (WSJ $)
- Amazon has confirmed the loss of three data centers in the strikes. (Business Insider $)
- GPS spoofing and jamming attacks have surged, disrupting regional maritime navigation. (Wired)
- Cryptocurrency markets are selling off sharply in response to the escalating uncertainty. (Bloomberg)
2 Apple is reportedly in talks to integrate Google's Gemini AI into Siri
The move would also deepen Apple's dependence on Google's cloud infrastructure at a pivotal moment for on-device AI. (The Information $)
3 A new database maps which subjects are being suppressed under the Trump administration
National parks are under pressure to remove exhibits deemed to contain "partisan ideology." (WP $)
- A transatlantic confrontation over free speech online is taking shape. (FT $)
- A firsthand account of being barred from the US for combating online hate speech. (MIT Technology Review)
4 Can AI genuinely augment workers rather than displace them?
Three economists make the case for cautious optimism. (New Yorker)
- A framework for steering AI development toward broadly shared economic prosperity. (MIT Technology Review)
5 Workplace surveillance software is growing more sophisticated by the day
So-called "bossware" tools are evolving well beyond basic activity monitoring. (NYT)
- Hiring managers are increasingly bypassing resumes altogether. (Business Insider)
6 RFK Jr. signals plans to lift the FDA ban on 14 regulated peptides
The move would reverse a Biden-era restriction on their compounding and distribution. (Gizmodo)
- Peptides are appearing everywhere in health and biotech — here's what they actually are. (MIT Technology Review)
7 Meta is piloting an AI-powered shopping research assistant
The tool is positioned as a direct challenger to Gemini and ChatGPT in the consumer commerce space. (Bloomberg)
8 Orbital data centers may be less far-fetched than the industry assumes
With the right enabling technologies, the economics could actually work. (Economist)
- A serious look at whether relocating data center infrastructure to space is worth pursuing. (MIT Technology Review)
9 Climate change is making air turbulence measurably more severe
Atmospheric scientists say the trend is accelerating — and passengers will feel it. (New Yorker)
10 6G is already generating hype — and the standards aren't even finalized yet
The industry's familiar hype cycle is spinning up right on schedule. (The Verge $)
Quote of the day
"We don't list markets directly tied to death. When there are markets where potential outcomes involve death, we design the rules to prevent people from profiting from death."
—Tarek Mansour, CEO and founder of prediction market platform Kalshi, attempting to defend the existence of a $54 million speculative position on "Ali Khamenei out as Supreme Leader?" listed on his platform, 404 Media reports.
One More Thing
South Africa's privatized surveillance apparatus is engineering a digital apartheid
Johannesburg has become the unlikely incubator of a distinctly South African model of mass surveillance. Over the past decade, the city has developed a centralized, coordinated, and entirely privately operated surveillance infrastructure — one that the local security industry has embraced with considerable enthusiasm, driven by the persistent pressures of operating in a high-crime urban environment.
Civil liberties advocates warn that this expanding surveillance apparatus is entrenching a new form of digital apartheid, systematically eroding the democratic rights that South Africans fought to secure. But a growing body of researchers argues the implications extend even further.
These scholars contend that the deployment of artificial intelligence is replicating the structural patterns of colonial governance — and that South Africa, where the legacies of colonialism remain deeply embedded, offers a particularly stark case study. The unchecked rollout of AI-driven surveillance here illustrates a troubling paradox: a technology marketed as a vehicle for social progress that is instead threatening to resurrect the hierarchies of the past. Read the full story.
—Karen Hao and Heidi Swart
We can still have nice things
A place for comfort, fun and distraction to brighten up your day. (Got any ideas? Drop me a line or skeet 'em at me.)
+ A growing cohort of digital creators is leveraging their online platforms to champion the survival of Britain's traditional pubs — turning social media reach into a preservation movement for one of the country's most iconic cultural institutions.
+ Ever wondered what the American continent looks like stripped down to its bare hydrological skeleton? This river-only map renders the US exclusively through its waterways, revealing a surprisingly intricate and beautiful network that most people never think about.
+ The Underwater Photographer of the Year has been announced — and this year's winning shot is as charming as it is technically stunning.
+ Three decades in, Pokémon shows no signs of slowing down. The franchise marks its 30th anniversary with a collector's market and fan base that have only grown more fervent with time.
