Authored by Jon Fleetwood via JonFleetwood.com,
In 2018, the Defense Advanced Research Projects Agency (DARPA) proposed building a portable computer system allegedly capable of determining whether a person had been exposed to biological or chemical threats - including weapons of mass destruction (WMD) and their precursors.
Even with no physical evidence.
DARPA outlined the proposal in Broad Agency Announcement HR001118S0023 for a program known as ECHO (Epigenetic CHaracterization and Observation).
DARPA proposed accomplishing this by comparing purported changes inside the person's own biology against a government-built reference library and algorithmically identifying both the alleged threat and when the exposure supposedly occurred.
Viewed alongside DARPA's broader pandemic-era research portfolio, the proposal also raises a larger question explored by this publication's In Silico Theory: as governments increasingly replace direct observation with computational models, reference libraries, and algorithmic determinations, where does objective evidence end and computer-generated conclusions begin?
It also raises an immediate practical question.
If an algorithm operating on a government-controlled reference library determines that someone was exposed to a dangerous biological or chemical threat without physical evidence, what meaningful mechanism exists for independently verifying - or successfully challenging - that determination?
SAM.gov search results associated with DARPA solicitation HR001118S0023 identify at least $53.8 million in awards, including a $49.2 million contract awarded to the Icahn School of Medicine at Mount Sinai and a $4.5 million contract awarded to Arizona State University.
The Core CapabilityRather than searching for the biological or chemical agent itself, ECHO proposed "identifying" alleged long-lasting biological signatures left behind after an exposure.
DARPA described the objective:
"utilize an individual's epigenome to reveal their history of exposure to weapons of mass destruction (WMD) and WMD precursors."The solicitation explains why DARPA believed this was possible:
"the epigenome is persistent and detectable even when physical evidence has been erased."That sentence is arguably the centerpiece of the entire program.
The proposed system was designed to continue generating exposure determinations long after the original biological or chemical material was no longer available for direct examination.
Epigenetics is the study of purported chemical modifications associated with DNA that researchers say can change in response to biological and environmental influences without altering a person's genetic code.
A Government-Built Exposure LibraryTo accomplish that objective, DARPA instructed performers to create what it called the ECHO Signature Panel (ESP) - a centralized library containing purported epigenetic signatures associated with specific exposures.
Contractors were required to generate signatures for a minimum of 21 exposure conditions, including mandatory biological threat agents and additional chemical, radiological, explosive, and related materials.
Future unknown samples would then be compared against that growing reference panel.
DARPA explained the process:
"The developed algorithms will need to compare an unknown exposure against the ESP to identify the specific exposure, and when the individual was exposed."In other words, the proposed system would not simply measure biology.
It would compare an individual's biological patterns against a government-built reference library and output a determination regarding both the nature and timing of an alleged exposure.
The Computer Makes The DeterminationThe solicitation envisioned an almost completely automated process.
DARPA required:
"no human intervention between the time the sample is introduced and results pertinent to exposure and timing is provided."Insert the sample.
Run the algorithms.
Receive the determination.
The individual would not be interpreting raw laboratory findings.
The system itself would produce the conclusion.
When The Person Becomes The EvidenceTraditional forensic investigations often rely on recovering the biological or chemical agent itself.
ECHO proposed something different.
According to the solicitation, the system was intended to continue identifying exposure after the physical evidence had been erased by relying instead on computational comparison against stored biological signatures.
That architecture raises an obvious question.
If the original biological or chemical material no longer exists, and the determination instead comes from an algorithm comparing an individual's biology against a government-controlled reference library, how would an ordinary person independently verify - or meaningfully challenge - that conclusion?
Could the system be weaponized or used fraudulently?
The document itself does not answer that question.
Instead, it focuses on improving the speed, automation, and reliability of the computational determination.
DARPA Anticipated Attempts To Defeat The SystemOne small line buried within the solicitation is especially revealing.
Among the categories requiring special handling, DARPA listed:
"Potential procedure to evade ECHO Signature Panel identification."The agency was not merely interested in creating exposure signatures.
It was already contemplating how those signatures - and the identification system built around them - might be circumvented.
Bottom LineMuch of the discussion surrounding ECHO focuses on epigenetics.
That risks obscuring what the solicitation actually proposed.
At its core, ECHO described a government-built computational system designed to determine whether an individual had been exposed to specified biological or chemical threats by comparing representations of that person's biology against a centralized signature library - even after the underlying physical evidence was gone.
Whether such a system could perform as intended is a separate scientific question.
The more immediate journalistic question is what follows if such algorithmic determinations are relied upon in military, intelligence, public-health, or legal settings.
When there is no physical evidence and the conclusion rests on a proprietary reference library and algorithmic comparison, what avenues exist for independent verification or meaningful challenge?
What happens to the person after the algorithm determines they've been exposed?
And if those algorithmic determinations are ever used to justify quarantine, criminal investigation, military action, employment decisions, travel restrictions, or other government action, what protections exist for the person the computer has already declared was exposed?
Follow Jon Fleetwood on Instagram and X, and subscribe at JonFleetwood.com. (h/t Modernity.news)
Tyler Durden Wed, 07/22/2026 - 17:40Nate Silver identifies lower-income, highly educated voters as the core demographic of the Democratic Socialists of America, revealing a striking breakdown in the traditional relationship between education and higher income.
"Income and education are usually positively correlated. But sometimes the relationship breaks down. What's the most Democratic voting group in America? People with post-graduate degrees but lower-to-middle incomes," Silver wrote on X.
In today's newsletter, I argue this is the sweet spot for the DSA. DSA members are *very* college-educated (80% BA+) but *not* particularly well-off (45% HHI incomes below $60K). Their class status is ambiguous. But they tend to be very liberal on culture as well as economics. pic.twitter.com/yXAW5pVczi
— Nate Silver (@NateSilver538) July 21, 2026One explanation: newly minted graduates burdened by student debt, holding low-value liberal arts and other worthless woke degrees, and working multiple service-sector jobs. The result is a cohort driven by progressive ideology but denied the prosperity they were promised, creating fertile recruiting ground for far-left movements.
Oh my lord.
A student went off the rails because a police officer used the “wrong pronouns.”
Officer: “He’s having a breakdown.”
Student: “SHE!!!!! SHE!!!!!! JUST SAY THE RIGHT PRONOUN!” (screaming like a lunatic)
We have a serious mental health crisis in this country, and… pic.twitter.com/kMF5GXfWB5
Some of these movements, including ones pushed by the DSA, aim to collapse the nation from within.
Whether intentional or not, an education system increasingly influenced by Marxism is producing politically radical graduates with insurmountable debt, primed to demand sweeping economic change. It's like the kids are being set up for failure even before stepping into a job.
By contrast, voters with high school or partial college educations who secure well-paying, goods-producing jobs tend to lean more Republican. These folks don't go through four to six years of woke indoctrination camps.
Palantir CEO Alex Karp on Zohran Mandani recently said it best about educated socialist voters:
The average Ivy League grad voting for this mayor is annoyed their education is not that valuable—and that the person who knows how to drill for oil has a more valuable profession. That annoys the f*ck out of these people.
Palantir CEO Alex Karp on Zohran Mandani voters:
“The average Ivy League grad voting for this mayor is annoyed their education is not that valuable—and that the person who knows how to drill for oil has a more valuable profession.”
“That annoys the fuck out of these people.” https://t.co/46r7RJLD9U pic.twitter.com/GG689vn4rk
The education system run by the left wing has massively failed young people. It is instead transforming them into Marxist activists intent on overthrowing America and capitalism - a trend evident in the rise of the DSA.
There is good news: Nancy Lazar, Piper Sandler's chief global economist and head of the firm's economics research team, recently published a note stating that goods-producing jobs are returning due to reshoring and data center buildouts.
For parents focused on positioning their children to thrive, the traditional college pathway can no longer be treated as the default. Unless a degree provides a clear return on investment, technical training, skilled trades or direct entry into goods-producing industries may offer a stronger path to financial independence.
Otherwise, why subject your kids to a four-year Marxist indoctrination camp, only for them to dye their hair purple, become weekend No Kings activists and work two jobs, one at Starbucks and the other as a cat walker?
Tyler Durden Wed, 07/22/2026 - 17:20Update (5:00pm ET): It took just a few minutes into the call for management to get to the elephant in the room, namely the relentless spending on overpriced commodity chips, and just like last quarter when GOOGL guided to a much higher capex for the current year, it did so again (if to a slightly lesser extent). The stock promptly slumped 3% to new session lows.
And then:
There was more:
In kneejerk reaction the stock promptly slumped 4% in AH trading to new session lows.
* * *
Earlier:
As we wrote in our GOOGL preview, according to UBS and JPM analysts, the market would be watching two things above all else when Alphabet became the first hyperscaler to report earnings after the close today: capex guidance and monetization language, as "the debate has shifted from "how big is capex" to "who proves monetization and returns." And yet, in light of the massive expectations for capex growth, the company was facing "downside risk" it it disappointed the buyside bars which were $100BN higher ($350-$375BN) than the median sellside estimate of $250BN.
So what did GOOGL just report for fiscal second quarter moments ago? Well, Alphabet reported Q2 cloud revenues that surpassed Wall Street’s expectations but slightly missed on sales tied to its search-engine business, threatening to fuel concerns about its heavy spending on artificial intelligence.
The Google parent also said cloud sales totaled $24.77 billion for the quarter ended June 30, a jump of 82% over the same period last year. That was better than the $22.46 billion that analysts expected, but well short of some buyside whisper bogeys. Sales from search advertising generated $63.27 billion, just below estimates for $63.28 billion.
Here are the details:
Here is what the company's explosively growing capex looks like now:
Some other notable highlights from the report:
Commenting on the quarter, CEO Sundar Pichai said that “our popular AI features are driving Search query growth” although many would counter that GOOGL's popular AI features were popular 9 months ago and users havelong since moved on to other platforms. He added that “Google Cloud saw a meaningful acceleration in growth as revenues increased 82% to $24.8 billion, led by an increase in Google Cloud Platform (GCP) across enterprise AI Solutions and enterprise AI Infrastructure, as well as core GCP services.”
The company also said that Gemini models now process 22 billion API tokens per minute: “We are seeing strong demand for our security solutions, and our new Gemini 3.5 Flash Cyber delivers highly cost- efficient performance at the frontier.”
Amusingly, the company said that it still hasn’t sold any shares under ATM program, a program which otherwise would be reserved to C-grade meme stonks desperate for liquidity. Come to think of it, if GOOGL plans on unveiling some insane capex number like $400BN or so for 2027, it will absolutely be desperate for liquidity in just a few months.
Alphabet is the first major US hyperscaler to report earnings this season, offering an early glimpse of what’s to come. The tech sector has committed hundreds of billions of dollars to data centers and other AI offerings, but has yet to show any measurable returns for the trillions it plans on spending on commodities like GPUs and extremely overpriced memory chips. Google is set to spend more than ever on capital expenditures this year to compete in the AI race, and investors have been waiting to see whether that spending is driving new growth or ultimately weighing on profitability. And it appears that investors are starting to get cold feet waiting for the ROI and margin inflection point (which according to Goldman's agentic propaganda report was supposed to take place this past quarter... it didn't).
The lukewarm results threaten to intensify investor scrutiny of Alphabet’s AI spending, which is projected to reach record highs in 2026. Wall Street has been scouring for evidence that the company’s massive investments are generating new growth rather than weighing on profits. As the first major US technology company to report earnings this season, Alphabet’s results also stand to shape expectations for a sector that has collectively committed hundreds of billions of dollars to AI infrastructure.
Alphabet shares slipped about 1% in after-hours trading Wednesday, after earlier swining to wider losses and modest gains. The stock closed at $342.09 per share in New York, and is essentially unchanged since early February.
The tepid share reaction despite the cloud beat, which offsets the search miss, reflects concerns over what the company will say about future capex and into 2027. During its last earnings call, management said they expect it to “significantly increase” over 2026; if they do the same this time, the market may not be very pleased as the ROI is still lacking.
Tyler Durden Wed, 07/22/2026 - 17:10