In partnership with
October 21, 2026 · READ ONLINE
Keeping you ahead in AI.
THE DAILY BRIEFING / 4 MIN READ
Welcome back, AI enthusiasts.
Anthropic is spending billions to stay ahead in AI while promising to put safety first. Now, as CEO Dario Amodei defends that promise against growing criticism, the company's latest financial disclosures reveal just how much is at stake. Today's story looks at whether those two priorities can hold together as the AI race gets more expensive.
IN TODAY'S ISSUE
01 Anthropic's CEO Defends Safety Amid $80 Billion Loss
02 OpenAI and Anthropic Race Into Drug Development
03 Claude Makes a Breakthrough in Scorpion Venom Research
01 / AI SAFETY
Evolving AI: Anthropic's safety commitments are facing scrutiny as new IPO disclosures reveal the financial stakes behind its AI ambitions.
KEY POINTS
Anthropic reported an $80.7 billion net loss in the first half of 2026, largely due to an accounting charge.
AI safety advocate Nathan Calvin raised nine concerns about Anthropic's priorities, prompting a detailed defense from CEO Dario Amodei.
Anthropic has committed at least $518 billion to computing and infrastructure over the next decade.
DETAILS
Anthropic CEO Dario Amodei defended the company's safety strategy after Calvin questioned whether its push for stronger AI was overtaking its original mission. Calvin pointed to loosened safety commitments and a controversial draft federal AI bill. Amodei acknowledged mistakes but argued that staying at the AI frontier helps raise industry safety standards, citing restricted model releases. Meanwhile, $72 billion of Anthropic's reported loss came from an accounting charge, while revenue reached $16.3 billion in the first half of 2026, up from $1.1 billion a year earlier.
WHY IT MATTERS
Anthropic's growing business depends on releasing stronger AI models to justify its enormous computing investments and sustain revenue growth. That makes Dario Amodei's defense of its safety mission especially important as the company prepares for its IPO. He argues that staying at the frontier gives Anthropic the influence to push the industry toward safer AI, and the company has made real decisions to restrict powerful models. But with OpenAI, Google, and other rivals competing for the same customers, delaying a release could mean losing business and falling behind. As the financial stakes grow, Anthropic's biggest test will be whether it can hold back a commercially important model over serious safety concerns, even when investors, customers, and competitors are pushing it forward.
Blu Dot surpasses 2,000% ROAS with self-serve CTV ads
Home furniture brand Blu Dot blew up on CTV with help from Roku Ads Manager. Here’s how:
After a test campaign reached 211,000 households and achieved 1,010% ROAS, the brand went all in to promote its annual sales event. It removed age and income constraints to expand reach and shifted budget to custom audiences and retargeting, where intent was strongest.
The results speak for themselves. As Blu Dot increased their investment by 10x, ROAS jumped to 2,308% and more page-view conversions surpassed 50,000.
“For CTV campaigns, Roku has been a top performer,” said Claire Folkestad, Paid Media Strategist, Blu Dot. “Comping to our other platforms, we have seen really strong ROAS… and highly efficient CPMs, lower than any other CTV partner we've worked with.”
Using Roku Ads Manager, the campaign moved from a pilot to a permanent performance engine for the brand.
02 / BIOTECH AI
Source: Ashoka Rajendra
Evolving AI: OpenAI has acquired Orchestra Bio to expand its Rosalind platform into managing drug development, bringing AI-powered coordination alongside its scientific research tools.
KEY POINTS
Orchestra's AI agents monitor drug studies and identify how setbacks, such as a failed manufacturing batch, could affect clinical trials and timelines.
OpenAI plans to connect Orchestra with GPT-Rosalind and Rosalind Workbench, which help scientists analyze biological data, explore research questions, and plan experiments.
Developing a new medicine takes roughly 10 to 15 years on average in the U.S., a lengthy process OpenAI hopes to speed up by reducing operational delays.
DETAILS
OpenAI has acquired Orchestra Bio, a startup that built an AI-powered platform to coordinate drug development across research teams, manufacturers, and outside partners. Its AI agents help manage complex workflows, track progress, and respond to setbacks that could delay clinical trials. OpenAI plans to combine these capabilities with Rosalind's scientific models and research tools. Meanwhile, Anthropic is also expanding into biotech through Claude Science and its own acquisitions.
WHY IT MATTERS
OpenAI and Anthropic are competing to become major AI providers for pharmaceutical companies. Anthropic reportedly acquired Coefficient Bio for $400 million, launched Claude Science, and established its own biology lab. With companies like Novo Nordisk already partnering with both, this rivalry could bring drugmakers better tools, more choices, and lower costs. But specialist biotech AI startups face growing pressure as frontier model providers expand into their territory, forcing them to prove why their software is worth paying for.
TOGETHER WITH DATADOG
Evolving AI: Why you should evaluate every AI agent before shipping to production, and how.
If you're shipping AI agents without offline validation, your users are the ones doing the testing. Get the free practical framework for evaluating production-grade AI agents before they hit production.
Get the guide to learn how to:
Build annotated test datasets that cover core use cases, edge cases, and adversarial inputs
Design deterministic and LLM-as-a-judge evaluators that reflect real business impact
Trace multi-agent workflows end-to-end during experimentation to catch failures before users do
Prevent model drift by keeping your offline test environment aligned with production
03 / AI BREAKTHROUGH
Source: Anthropic
Evolving AI: Anthropic's Claude designed small proteins to target dangerous scorpion toxins, with real laboratory tests confirming that some successfully attached to them.
KEY POINTS
In just 64 hours, Claude used 12 specialized tools to generate protein designs, guided by one initial prompt and four follow-ups.
Of 211 designs tested in laboratories, 17 successfully bound to four of the five targeted scorpion toxins.
Scorpions sting over 1.2 million people annually, causing more than 3,250 deaths, while existing antivenoms are often expensive or difficult to access.
DETAILS
Anthropic teamed up with investment research firm Citrini Research to see whether Claude could help develop a new kind of scorpion antivenom. Claude used specialized protein-design tools to create and screen tiny proteins intended to attach to venom toxins before they disrupt nerve signals. Adaptyv Bio and Twist Bioscience then tested the candidates in the lab. The next challenge is proving these proteins can actually neutralize venom in living cells and animals.
WHY IT MATTERS
Claude’s experiment hints at a future where sophisticated drug research becomes accessible to smaller teams with limited scientific expertise. That could be especially valuable for neglected medical problems like scorpion stings, where millions are affected but commercial incentives to develop new treatments are often weak. If AI can make early drug discovery faster and cheaper, more researchers could explore treatments that might otherwise never receive funding. However, turning promising designs into actual medicines still requires extensive testing, specialized labs, and significant investment. As AI takes on more of the research process, the biotech companies that can validate and develop these discoveries could become increasingly important.
04 / HUMAN OR MACHINE?
One is real. Trust your eyes?
Two images. One photograph. Click the one you think is real.
Image 1
Image 2
05 / QUICK HITS
💳 Mastercard opened Agent Pay for Machines for developer testing after an AI agent made nine payments within one approved budget.
🚨 OpenAI disclosed that two internal AI agents weakened safety monitoring during coding tasks, with one falsely claiming it had not removed a monitor.
🚀 SpaceX reported $7.56B in quarterly AI revenue, nearly triple the previous quarter and more than its Starlink connectivity business generated.
🚗 Google is bringing Gemini's Personal Intelligence to cars, letting drivers retrieve flight details, appointments and saved places through connected Google apps.
06 / TRENDING AI TOOLS
🚀 Fyxer - email assistant that organizes your inbox, writes draft replies in your tone, and takes meeting notes*
🛠️ Floot MCP - Builds and ships full-stack apps inside Claude or ChatGPT.
🎬 Rushilo - Edit videos by telling Claude Code or Codex what to change.
🎤 Toby - Runs realistic mock interviews and tells you what to improve.
*PARTNER LINKS
BEFORE YOU GO
Worth your four minutes?
Help us make the next edition better.
See you tomorrow,
The Evolving AI team
