Snap Loses Top Specs Executive

OpenAI builds AI devices for 2027 launch, the Supreme Court strikes down sweeping Trump tariffs, and SoftBank unveils $33B Ohio power plant.

Welcome back to your daily memorandum talking tech, business, AI, markets, and more. đŸ—žïž

In today’s edition, we are tackling the following:

đŸ€– OpenAI is assembling a 200-person team for an AI hardware push.
⚖ Supreme Court voids Trump tariffs, triggering tech stock rally.
đŸ”„ SoftBank plans $33B Ohio plant powering AI data centers.
đŸ—łïž Venezuelan opposition leader freed after sweeping amnesty law.
🔎 Prediction markets embed betting odds into mainstream newsrooms.

What investment is rudimentary for billionaires but ‘revolutionary’ for 70,571+ investors entering 2026?

Imagine this. You open your phone to an alert. It says, “you spent $236,000,000 more this month than you did last month.”

If you were the top bidder at Sotheby’s fall auctions, it could be reality.

Sounds crazy, right? But when the ultra-wealthy spend staggering amounts on blue-chip art, it’s not just for decoration.

The scarcity of these treasured artworks has helped drive their prices, in exceptional cases, to thin-air heights, without moving in lockstep with other asset classes.

The contemporary and post war segments have even outpaced the S&P 500 overall since 1995.*

Now, over 70,000 people have invested $1.2 billion+ across 500 iconic artworks featuring Banksy, Basquiat, Picasso, and more.

How? You don’t need Medici money to invest in multimillion dollar artworks with Masterworks.

Thousands of members have gotten annualized net returns like 14.6%, 17.6%, and 17.8% from 26 sales to date.

*Based on Masterworks data. Past performance is not indicative of future returns. Important Reg A disclosures: masterworks.com/cd

*This is sponsored content. See our partnership options here.

TECHNOLOGY

  • Inside the OpenAI team developing AI devices (The Information) 

    More: 9to5Mac, The Decoder

    • OpenAI has assembled a 200-person hardware team building a family of AI devices, with a smart speaker planned as the first launch.

    • The speaker, priced between $200 and $300, will include a camera and facial-recognition purchasing feature; shipping no earlier than February 2027.

    • The lineup also includes smart glasses (not ready for mass production until 2028), a smart lamp, an audio wearable codenamed "Sweetpea," and a stylus called "Gumdrop."

  • Multiple AWS outages caused by AI coding bot blunder (Tom's Hardware) 

    More: Financial Times

    • Amazon's internal AI coding tool Koiro erased its own working environment in December, triggering a 13-hour AWS outage in parts of mainland China.

    • A second incident had no customer-facing impact; in both cases, engineers gave the AI agent unrestricted permissions with no secondary approval required.

    • Amazon insists both failures were "user error, not AI error," but has updated internal controls to prevent autonomous AI agents from taking destructive actions unsupervised.

  • Nvidia, Dell, and Apple rally as SCOTUS strikes down Trump tariffs (Barron's) 

    More: CNBC, Bloomberg, NPR

    • The Supreme Court struck down Trump's sweeping "Liberation Day" tariffs 6-3, ruling he exceeded his authority under the International Emergency Economic Powers Act.

    • Tech stocks led the rally: Nvidia, Dell, and Apple surged as investors priced in lower component and assembly costs, with the Nasdaq jumping over 1%.

    • The ruling is silent on $175B in potential refunds to importers, and the Trump administration vowed to pursue tariffs via alternative legal avenues.

BUSINESS

  • At a critical moment, Snap loses a top Specs exec (TechCrunch) 

    More: Road to VR, Reuters

    • Scott Myers, SVP of Specs and the top hardware executive behind Snap's upcoming consumer AR glasses, has left the company following a reported dispute with CEO Evan Spiegel.

    • His departure comes just weeks after Snap spun its AR team into a standalone subsidiary, Specs Inc., and less than a year before the glasses' planned consumer launch.

    • Snap disputed the "blow-up" characterization but offered no additional explanation, saying it "can't wait to bring Specs to the world later this year."

  • General Catalyst's Hemant Taneja on $5B India bet (Economic Times) 

    More: TechCrunch, YourStory, Bloomberg

    • General Catalyst announced a $5B commitment to India over five years at the India AI Impact Summit, a five-fold increase over its previous India allocation.

    • CEO Hemant Taneja said the firm sees India's AI opportunity in large-scale real-world deployment —not frontier model building—across AI, healthcare, fintech, and defense.

    • The firm is shifting from traditional VC to a "company creation" model, aiming to build and scale startups alongside investing, with India as its largest international bet.

  • SoftBank to form consortium for $33B Trump deal power plant (Nikkei Asia) 

    More: TechCrunch, Power Magazine, Reuters

    • SoftBank subsidiary SB Energy will build a 9.2-gigawatt natural gas power plant in Portsmouth, Ohio—part of a broader $36B Japan–US investment package announced by Trump.

    • The $33B project would be the largest natural gas plant in US history, capable of powering approximately 7.5M homes, with groundbreaking scheduled for late 2026.

    • The plant is intended to supply baseload power to AI data centers, including those tied to SoftBank's Stargate initiative with OpenAI and Oracle.

  • Stay ahead of AI at work*

    • Cut through the hype. Roko’s Basilisk distills the day’s AI breakthroughs, market moves, and workable workflows into a five-minute brief.

    • Get one sharp deep dive, fast Quick Hits, and a witty pro tip, plus links that actually matter.

    • Join 100k+ readers who start decisions here each morning.

MARKETS

S&P

6,861.89

−0.28%

NASDAQ

22,682.70

−0.31%

Dow

49,395.20

−0.54%

10-Year

4.07%

↓ ~0.01 pp

Bitcoin

67,765

+0.01%

Gold

5,037

+0.79%

Stay ahead of AI at work.

Cut through the hype. Roko’s Basilisk distills the day’s AI breakthroughs, market moves, and workable workflows into a five-minute brief.

Get one sharp deep dive, fast Quick Hits, and a witty pro tip, plus links that actually matter.

Built for builders, operators, and curious leaders who need context fast. Join 100k+ readers who start decisions here each morning.

*This is sponsored content

WORLD

  • Trump tariffs ripped up the global trade order. What now? (BBC) 

    More: CNBC, Bloomberg, OPB/NPR

    • The Supreme Court struck down Trump's IEEPA-based "Liberation Day" tariffs in a 6-3 ruling, ending the broadest presidential tariff experiment in nearly a century.

    • The US had been collecting roughly $30B a month in tariff revenue; up to $175B in refunds to importers could now be owed, though the majority opinion was silent on repayment.

    • The administration called the decision a "disgrace" and immediately signaled it would rebuild tariffs through Section 232 national security powers and other legal routes.

  • Venezuelan opposition politician released after amnesty law passed (BBC) 

    More: Al Jazeera, Bloomberg, PBS

    • Juan Pablo Guanipa, a senior Venezuelan opposition figure and close Machado ally, announced his release after "almost nine months of unjust imprisonment" following the amnesty law's passage.

    • Venezuela's legislature approved the Amnesty Law for Democratic Coexistence covering political offenses since 1999, potentially freeing more than 600 detainees targeted over 27 years of Chavista rule.

    • Guanipa immediately criticized the law as a "flawed document," citing exclusions and judicial restrictions that leave many prisoners still behind bars; some rights groups expressed cautious optimism.

  • Samba school that praised Brazil's Lula at Rio Carnival relegated (BBC) 

    More: France 24, AP / Washington Post, Merco Press

    • AcadĂȘmicos de NiterĂłi, which honored President Lula with Rio's first-ever tribute parade to a sitting president, finished last in the Grupo Especial rankings and was relegated to the second division.

    • The parade ignited a political firestorm after a segment depicted the "traditional family" inside a tin can, which conservatives framed as an attack on Christian values; Bolsonaro's allies celebrated the relegation.

    • With Brazil's presidential election in October, FlĂĄvio Bolsonaro mocked the result, saying the "next drop" will be Lula's—while opposition parties have filed electoral court complaints alleging illegal early campaigning.

FUTURISM

  • Prediction markets are buying journalism. What does that mean? (The Verge) 

    More: Axios, The Intercept, NPR

    • Prediction markets Kalshi and Polymarket have signed sponsorship and partnership deals with CNN, CNBC, Dow Jones, Yahoo Finance, and Substack, embedding gambling odds directly into mainstream news coverage.

    • Both platforms have spread false claims—including a fabricated US-Denmark negotiation over Greenland and an Epstein-era fake Jeff Bezos quote—with few editorial guardrails applied to their social media accounts.

    • Critics warn that insider trading is structurally unchecked: hours before the US raid on Maduro, one anonymous trader pocketed $400k on a suspiciously well-timed Polymarket bet.

  • The Claude C compiler: what it reveals about the future of software (Modular) 

    More: Anthropic Engineering Blog, VizOps / agent scaling laws

    • Anthropic's Claude built a full C compiler in Rust—with a frontend, optimizer, and backends for x86, ARM, and RISC-V—demonstrating that AI can now maintain architectural coherence across large engineering systems.

    • LLVM co-creator Chris Lattner calls it a real milestone, but notes it reproduces known textbook design patterns rather than inventing new abstractions, reflecting AI's strength in implementation over genuine innovation.

    • The post argues that as AI automates implementation, the scarce skills become architecture, design judgment, and community building—and that AI amplifies both good and bad structure at scale.

  • Amazon's shiny new warehouse robot just failed in spectacular style (Futurism) 

    More: Business Insider, TechCrunch, NYT

    • Amazon quietly killed its multi-arm Blue Jay warehouse robot just months after a splashy October announcement, with internal sources telling Business Insider the project was abandoned without public explanation.

    • A spokesperson said the underlying technology will be "repurposed" across other warehouse systems, though the company's October press release never described Blue Jay as a prototype.

    • The failure highlights the persistent gap between AI lab advances and physical-world robotics, even as Amazon commits to $200B in AI infrastructure spending this year and plans to eventually replace 600k jobs with robots.

CONTENT

EXTRAS

  • Ex-Google engineers charged with stealing phone processor tech (Bloomberg) 

    More: CNBC, DOJ Press Release, The Register

    • Two sisters—former Google hardware engineers Samaneh and Soroor Ghandali—and Samaneh's husband were indicted on 14 felony counts of conspiracy, trade secret theft, and destruction of evidence.

    • The trio, all Iranian nationals, allegedly exfiltrated hundreds of confidential files tied to Google's Tensor processor and Snapdragon-class SoC technology, with some data transferred to unauthorized locations in Iran.

    • Google detected the theft through routine security monitoring and referred the case to the FBI; the DOJ framed it as a national security threat targeting core American semiconductor innovation.

  • Perplexity's retreat from ads signals a shift in how AI search makes money (Wired) 

    More: Financial Times, TechSpot, Search Engine Land

    • Perplexity has fully wound down the sponsored answers it tested in 2024, with executives confirming to the Financial Times they have no plans to revive advertising on the platform.

    • The decision is a direct trust bet: the company argues that even labeled ads make users "doubt everything," undermining the accuracy-first identity that drives its 100M users and $200M in annualized subscription revenue.

    • The move is a striking contrast to OpenAI (which just began testing ChatGPT ads) and Google (which runs ads in AI Overviews); Anthropic has also committed to keeping Claude ad-free.

  • Google says its AI systems helped deter Play Store malware in 2025 (TechCrunch) 

    More: Google Security Blog

    • Google blocked 1.75M policy-violating apps from the Play Store in 2025, down from 2.36M in 2024, a decline it attributes to AI-powered deterrence raising the cost for bad actors.

    • The company banned 80k developer accounts attempting to publish malicious apps—also down sharply year-over-year—and now runs over 10k safety checks on every app published.

    • Meanwhile, Google Play Protect flagged more than 27M malicious sideloaded apps in 2025, nearly double 2024's total, suggesting bad actors are increasingly avoiding the Play Store and targeting users through unofficial channels.

AND MORE

P.S. Want to collaborate?

Here are some ways.

  1. Share today’s news with someone who would dig it. It really helps us to grow.

  2. Let’s partner up. Looking for some ad inventory? Cool, we’ve got some.

  3. Deeper integrations. If it’s some 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?

Login or Subscribe to participate in polls.