- Memorandum
- Posts
- Google Launches Gemini 3.1 Flash Lite
Google Launches Gemini 3.1 Flash Lite
Apple unveils M5 MacBook Pro models, OpenAI amends Pentagon contract after backlash, and AWS data centers face disruptions in the Middle East.
Welcome back to your daily memorandum talking tech, business, AI, markets, and more. 🗞️
In today’s edition, we are tackling the following:
💻 Apple debuts M5 Pro and M5 Max MacBook Pros.
🛡️ OpenAI restricts Pentagon contract to block domestic surveillance uses.
🌍 AWS warns Middle East operations remain unpredictable after strikes.
🎧 Audible launches cheaper streaming tier to challenge Spotify.
⚖️ Legal experts challenge US designation targeting Anthropic.

LLM Feature Integration with Real APIs
Building LLM-powered features means handling real APIs, real users, and real edge cases. You need engineers who can do the glue work that makes AI actually usable—Python, JavaScript, and everything in between.
Full-stack LLM integration
40–60% cost savings
This is the kind of talent you get with Athyna Intelligence: vetted LATAM engineers working in U.S.-aligned time zones.
*This is sponsored content. See our partnership options here.

TECHNOLOGY
Google launches Gemini 3.1 Flash-Lite, delivering enhanced performance Google)
More: Artificial AnalysisGemini 3.1 Flash-Lite launches in preview today at $0.25/1M input tokens and $1.50/1M output — Google's cheapest Gemini 3 model yet.
Outperforms Gemini 2.5 Flash with 2.5x faster time-to-first-answer and a 45% increase in output speed per Artificial Analysis benchmarks.
Scores 1432 Elo on Arena.ai leaderboard, 86.9% on GPQA Diamond, and 76.8% on MMMU Pro, surpassing several older, larger models.
Apple introduces MacBook Pro with all-new M5 Pro and M5 Max (Apple Newsroom)
More: MacRumors, 9to5Mac, Macworld
New 14- and 16-inch MacBook Pro models with M5 Pro and M5 Max chips launch today, starting at $2,199; pre-orders open on March 4, with shipping on March 11.
Built on a new dual-die "Fusion Architecture," the chips deliver up to 4x faster AI performance than M4 Pro and M4 Max, with Neural Accelerators in every GPU core.
Base storage jumps to 1TB for M5 Pro models, and 2TB for M5 Max, with SSD speeds up to 14.5GB/s and Wi-Fi 7 via Apple's new N1 chip.
Audible launches a cheaper, $8.99/month streaming tier in the US (Bloomberg)
More: TechCrunch, 9to5Mac, Deadline
Amazon's Audible launches a new "Standard" plan at $8.99/month — $6 less than its existing Premium tier — offering one audiobook credit and an ad-free curated listening library.
Unlike Premium, Standard titles are only accessible while subscribed; the plan also bundles nearly 200 former Wondery+ titles now available as Audible Originals.
The move directly targets Spotify, whose audiobook listening hours rose 37% last year among 281M premium subscribers who have engaged with the format.
BUSINESS
OpenAI amended its DOD contract to ensure "the AI system shall not be intentionally used for domestic surveillance" (CNBC)
More: Axios, Fortune, TechCrunchAltman shared an internal memo on X confirming OpenAI is amending its Pentagon contract to include explicit anti-surveillance and anti-autonomous-weapons language after widespread backlash.
The updated contract now bars use of commercially acquired personal data for surveillance, a gap critics said left geolocation data and browser histories unprotected in the original.
Altman acknowledged the deal was "opportunistic and sloppy" and said OpenAI shouldn't have rushed to sign it hours after Anthropic's supply-chain-risk designation.
AWS warns of "unpredictable" Middle East operations after drones "directly struck" two UAE facilities (CNBC)
More: Data Center Dynamics, Reuters, 404 Media
AWS confirmed that drones directly struck two of its UAE data center facilities and targeted a third near Bahrain as part of Iran's retaliatory strikes against US and Israeli targets.
EC2, EBS, and RDS services in the AWS ME-CENTRAL-1 region (Dubai) were disrupted for customers after one availability zone lost power and a second went offline due to a related power issue.
Amazon warned customers in a service health dashboard update that operations across the Middle East region remain "unpredictable" until the security situation stabilizes.
Some data annotators in Nairobi say they view private videos from Ray-Ban Meta glasses (Svenska Dagbladet)
More: Futurism, 9to5Mac, The Decoder
A joint investigation by two Swedish newspapers found that video from Meta's Ray-Ban AI glasses is routed to Sama data annotators in Nairobi, who describe seeing footage of users undressing, in bathrooms, having sex, and handling bank cards.
Workers said they felt unable to refuse the work, and that users wearing the glasses appeared entirely unaware that their footage was being reviewed by overseas contractors.
Privacy lawyer Kleanthi Sardeli of NOYB warned that once data enters Meta's training systems, users lose all practical control over how it is used — and EU regulators are likely to investigate.
Explore the essential startup reading database*
Learning from the best is how you become the best. We've curated a list of top-notch articles, blogs, and readings that are pivotal for startup founders.
And here's the kicker: we've organized it all in a user-friendly Notion database for easy access and tracking. Double win!
As aspiring leaders and innovators, we use this list as a bridge to absorb the knowledge that has propelled others to success.
MARKETS
S&P | 6,791.10 | −1.31% |
|---|---|---|
NASDAQ | 22,237 | −2.23% |
Dow | 47,787.55 | −2.29% |
10-Year | 4.10% | ↑ ~0.05 pp |
Bitcoin | 67,907 | −0.02% |
Gold | 5,068 | −0.75% |

Outperform the competition.
Business is hard. And sometimes you don’t really have the necessary tools to be great in your job. Well, Open Source CEO is here to change that.
Tools & resources, ranging from playbooks, databases, courses, and more.
Deep dives on famous visionary leaders.
Interviews with entrepreneurs and playbook breakdowns.
Are you ready to see what’s all about?
*This is sponsored content

WORLD
US Defense Secretary Pete Hegseth's and Trump's actions against Anthropic have serious legal issues (Lawfare)
Law professors Endrias and Rozenshtein argue in Lawfare that the Pentagon's supply-chain-risk designation of Anthropic is legally indefensible on multiple independent grounds — any one of which could sink it in court.
The statute cited (10 USC 3252) is a defense procurement law that has never previously been applied to an American company, and Trump's government-wide order has no apparent statutory basis beyond a Truth Social post.
Anthropic has vowed to sue, and legal experts say the designation may be intentional political theater, designed to pressure without ever surviving judicial review.
'Fear is everywhere': BBC reports from Mexican city turned into war zone by drug cartel feud (BBC)
BBC correspondent Quentin Sommerville reports from Culiacán, where rival Sinaloa cartel factions are waging open warfare, leaving paramedics in body armor and city streets deserted after dark.
The civil war within the Sinaloa cartel was ignited by the 2024 US arrest of leader Ismael "El Mayo" Zambada, triggering an internal power struggle that has left civilians caught in near-daily crossfire.
The report lands as the killing of CJNG leader El Mencho raises fears of a similar spiral of violence in Jalisco and as Trump threatens direct US military action against cartels.
US warns of Iranian attack on Saudi oil city as Trump tells Tehran it's 'too late' to talk (BBC)
The US issued intelligence warnings to Saudi Arabia of a potential Iranian strike on Abqaiq or Ras Tanura — the oil infrastructure that processes roughly 7% of global crude supply — following US and Israeli strikes that killed Ayatollah Khamenei.
Trump declared it was "too late" for Iran to negotiate, as Tehran launched a wave of retaliatory drone and missile strikes against US-aligned targets across the region over the weekend.
Oil markets spiked on the warnings as traders assessed the risk of major supply disruption; the strikes have already damaged AWS data centers in the UAE and Bahrain.
FUTURISM
Meta workers say they're seeing disturbing things through users' smart glasses (Futurism)
More: The Verge
Meta's Ray-Ban AI glasses, which sold over 7M units in 2025, are routing user video to offshore contractors with little apparent oversight or user awareness.
Workers described footage so intimate — bathroom visits, undressing, sex — that some felt they were violating users' privacy simply by doing their job.
Users who enable Meta's live AI features have no option to opt out of data sharing — meaning consent is buried in terms of service, which almost no one reads.
Ars Technica fires reporter after AI controversy involving fabricated quotes (Futurism)
More: 404 Media
Condé Nast has terminated senior AI reporter Benj Edwards after a retracted Ars Technica article was found to contain AI-fabricated quotes falsely attributed to a real person.
Edwards said he used an experimental Claude Code-based tool while sick with a fever and accidentally incorporated paraphrased AI output into the story as if it were a real source quote.
Editor-in-chief Ken Fisher called it a "serious failure of standards"; the case is among the highest-profile instances of an AI-assisted error ending a journalist's career at a major outlet.
Harvard Professor Says AI Users Are Losing Cognitive Abilities (Futurism)
More: Harvard astrophysicist on UFO AI, Harvard AI science paper, Fast Company on Loeb AI views
Harvard professor Avi Loeb argues that heavy reliance on AI may erode cognitive engagement and critical thinking among users.
Loeb likens modern AI dependency to superficial tech fascination that masks deeper intellectual costs.
The piece frames AI adoption as cultural, not just technological, with long-term impacts on human intellect.
The episode traces Formula 1’s evolution from a chaotic racing series to a global sports powerhouse.
Bernie Ecclestone’s centralized control built the league’s fortunes while limiting team economics.
Liberty Media’s acquisition strategy professionalized operations, driving multibillion-dollar valuations for its teams and explosive audience growth.
'Uncapped with Jack Altman: Garry Tan, Harj Taggar, and Jared Friedman from YC'(Uncapped)
YC’s leadership reflects on how the accelerator evolved while its core mission remains unchanged.
They explain how AI is reshaping founder expectations and lowering barriers to building startups.
The group argues that execution still outweighs competition as the decisive factor for success.
'Dario's Choice and Anthropic's Future, Apple's AI Devices, Netflix Loses WBD' (Big Technology Podcast)
M.G. Siegler joins to unpack Anthropic’s Pentagon clash and OpenAI’s intervention.
The discussion explores Dario Amodei’s leadership decisions and AI’s expanding role in defense.
Apple’s AI hardware plans and Netflix’s collapsed WBD deal round out the episode.

Would you trust AI to handle your startup’s finances? |
*This is sponsored content

EXTRAS
Google unveils Google Home updates to fix many annoyances and idiosyncrasies (The Verge)
More: 9to5Google
Google Home's biggest update in years arrives with a new Live Search feature that lets Gemini answer natural-language questions about what a camera is currently seeing in real time.
The update also addresses long-standing user complaints about smart home quirks, including improved device grouping, faster automation response times, and a redesigned dashboard.
Google Home chief Anish Kattukaran confirmed the changes, positioning Live Search as a direct response to the demand for AI-assisted home monitoring without requiring third-party hardware.
US criticism over censorship met with disbelief and anger (Financial Times)
European officials and diplomats have reacted to Trump administration attacks on EU digital regulation with a mix of bewilderment and fury, sources tell the FT.
Several Big Tech lobbyists privately acknowledge that US accusations of censorship targeting the DSA and DMA are exaggerated, even as their companies publicly support Washington's position.
The transatlantic rift over tech regulation is deepening at a moment when the EU is also navigating fraught trade negotiations with a second Trump term.
Meta is testing a shopping research feature in its Meta AI chatbot on the web for select U.S. users (Bloomberg)
More: 9to5Mac
Meta is quietly rolling out a shopping research tool within its Meta AI web interface, allowing select U.S. users to ask the chatbot to compare products, surface deals, and recommend purchases.
The feature puts Meta into direct competition with OpenAI's ChatGPT shopping integrations and Google's Gemini shopping tools, as all three platforms race to own the AI-assisted commerce funnel.
Meta did not announce a broader rollout timeline, and the test is currently limited to a subset of U.S. users on the web version of Meta AI.
AND MORE
France intends to send anti-drone and anti-missile systems to Cyprus after a British base was struck by drones.
Stock futures plunged as the Middle East conflict rattled markets and oil prices surged.
Your brand deserves to be in front of people who actually make decisions. That’s Memo’s audience.*
A rare total lunar eclipse, a “blood moon,” lights the skies across much of the world tonight.
Skywatchers in India are set to see a total lunar eclipse coinciding with the Holi festival tonight.
Nintendo confirmed a new Indie World Showcase featuring updates from indie developers for the Switch platform.
Banks in several Indian states are closed today in observance of Holi and related festivals.
Gulf nations struggle to replenish interceptor missiles amid relentless Iranian drone and missile attacks.
Energy infrastructure in the Gulf is on alert after an Iranian Shahed drone strike ignited a major UAE oil terminal.
P.S. Want to collaborate?
Here are some ways.
Share today’s news with someone who would dig it. It really helps us to grow.
Let’s partner up. Looking for some ad inventory? Cool, we’ve got some.
Deeper integrations. If it’s longer-form storytelling you are after, reply to this email, and we can get the ball rolling.

What did you think of today's memo? |

