The Download: the “steroid olympics” and a safer Mythos

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.

The “steroid olympics” were a circus—and a window into our culture

—Amit Katwala

A couple of weeks ago, at a $50 million arena built in a casino parking lot in Las Vegas, I witnessed a libertarian thought experiment come to life. The inaugural Enhanced Games were the first sporting competition where participants were encouraged to take performance-enhancing drugs.

For supporters of the event, the Enhanced Games offered a glimpse of a future in which medical advances push the human race to new heights—and they never have to get old. As I watched the games unfold, two questions bounced around my head: were they right? And what does that mean for the rest of us?

Read the full story to understand the answers.

MIT Technology Review Narrated: a reality check on the AI jobs hysteria

Despite the growing hysteria over AI’s threat to white-collar jobs, there’s still scant evidence that the technology has had a large-scale impact on the labor market.

Analysis of US labor data shows that unemployment in occupations most exposed to AI is actually lower than in less-exposed jobs. There are also no signs that large numbers of workers are shifting from AI-threatened professions into supposedly safer manual-labor jobs.

It’s true that things aren’t great in the job market. But the reason isn’t simply the rise of AI.


—David Rotman

This is our latest story to be turned into an MIT Technology Review Narrated podcast, which we publish each week on Spotify and Apple Podcasts. Just navigate to MIT Technology Review Narrated on either platform, and follow us to get all our new content as it’s released.

The must-reads

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.

1 Anthropic has released a “safe” version of Mythos
It promises it has enough guardrails and user limitations to be safe. (BBC)
+ It has a price tag twice as high as the previous flagship system. (NYT $)
+ Anthropic previously claimed Mythos was too dangerous to release. (CNBC)
+ But critics suspect that was a marketing play. (Guardian)
+ Selective access has become a key strategy for AI labs. (Axios)

2 Seattle has banned new data centers for a year
It’s the largest US city to have passed such a moratorium.(Guardian)
+ Its biggest tech firm, Amazon, has tried to stop the ban. (The Verge)
+ The movement to stop data centers is growing. (NYT $)

3 Democratic senators are pushing for a military AI restriction law
They want a human commander to have the final say. (Gizmodo)
+ But humans in the loop in an AI war is an illusion. (MIT Technology Review)

4 SpaceX plans to launch space data center tests by late 2027
Orbital compute is central to the company’s growth pitch. (Reuters $)
+ It’s also shared new designs for its space data centers. (BI)
+ We’d need these four things to put them in orbit. (MIT Technology Review)

5 China has been accused of escalating AI espionage
A report claims Beijing is hacking tech firms to catch up with the US. (CNBC)
+ There are no winners in a US-China AI arms race. (MIT Technology Review)

6 The Trump family has made about $2.3 billion from crypto
While investors lost about the same amount. (Gizmodo)
+ The Trumps risked next-to-nothing on their crypto ventures. (Reuters $)

7 Apple isn’t launching Siri AI in the European Union
It’s blaming EU interoperability requirements. (The Verge)
+Brussels says Apple didn’t try to find a compliance solution. (Reuters $) 

8 China’s new drone rules have spooked its thriving industry 
Drone firms face new commercial barriers. (Financial Times $)
+ China’s drone sector leads the world. (NYT $)

9 A judge has cancelled a trial after finding both legal teams used AI
The case descended into GenAI tools arguing against each other. (404 Media)
+ Courts have been flooded with AI-generated lawsuits. (MIT Technology Review)

10 The dinosaur-killing asteroid created a thriving new ecosystem
Microscopic life flourished in the extended heat. (New Scientist $)

Quote of the day

“AI technologies today are designed by and for WEIRD societies—Western, educated, industrialized, rich, and democratic.” 

—Aditya Vashistha, an assistant professor at Cornell University, tells Rest of World why AI systems don’t serve global needs.

One More Thing

""

LAUREN SIMKIN BERKE


Why the definition of design might need a change

The word “design” once carried a far wider set of meanings than it does today. They ranged from the literal and material (like tracing) through the tactical (to contrive and achieve a goal) to the organizational and institutional—the “designation” of people and objects.

Over centuries, as designing became increasingly separated from making, that broader understanding faded. But now there is a growing case for reclaiming the word’s original sense: not just the search for a more beautiful shape, but the shaping of a more beautiful and sustainable world. 

Find out why we should retool the word “design.”

—Nicholas de Monchaux

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.)

+ This history of humanity’s search for alien life is fascinating.
+ Watch a damaged painting slowly return to life in this art restoration video.
+ Admire young stars across every stage of cosmic formation in this stunning space picture of the month.
+ Daredevil divers have captured the first-ever underwater footage of an adult great white in the Mediterranean Sea.

Deep and repetitive transcranial magnetic stimulation improves motor dysfunction after basal ganglia infarction: preliminary findings on efficacy and electrophysiological mechanisms

ObjectiveTo observe the therapeutic effects of deep transcranial magnetic stimulation (dTMS) and repetitive transcranial magnetic stimulation (rTMS) on upper and lower limb motor dysfunction in patients with basal ganglia infarction, and to preliminarily explore their underlying electrophysiological mechanisms.MethodsThirty patients with motor dysfunction secondary to basal ganglia infarction, hospitalized at the Affiliated Hospital of North Sichuan Medical College between October 2024 and December 2025, were enrolled in this study. All eligible participants were randomly assigned to one of three treatment groups: dTMS (n = 10), rTMS (n = 10), or sham stimulation (n = 10). All patients in the three groups received routine medical treatment and conventional rehabilitation training. On this basis, the dTMS group was treated with 10 Hz dTMS, the rTMS group with 10 Hz rTMS, and the sham stimulation group with sham stimulation, 5 sessions per week for 2 consecutive weeks. Before treatment, on the first day after treatment, and at 30 days after treatment, the Fugl-Meyer Assessment (FMA), Berg Balance Scale (BBS), and Modified Barthel Index (MBI) were used to evaluate motor function of the affected side and activities of daily living. The resting motor threshold (rMT) and central motor conduction time (CMCT) of the affected hemisphere were measured simultaneously.ResultsThe baseline data among the three groups were comparable (all p > 0.05); After treatment, there was a statistically significant interaction between group and time in FMA-UE, FMA-LE, MBI, and BBS scores among the three groups (all p < 0.05); Compared with baseline, FMA-UE, FMA-LE, MBI, and BBS scores were significantly increased on the first day and at 30 days after treatment in all three groups (all p < 0.001); Compared with the sham stimulation group, the dTMS group exhibited higher FMA-UE, FMA-LE, MBI, and BBS scores on the first day and at 30 days after treatment (all p < 0.05); Compared with the rTMS group, the dTMS group showed no significant differences in FMA-UE and MBI scores on the first day after treatment (all p > 0.05), but higher FMA-LE and BBS scores (all p < 0.05), at 30 days after treatment, FMA-UE, FMA-LE, MBI, and BBS scores were all higher in the dTMS group (all p < 0.05). There was a statistically significant interaction between group and time in rMT and upper limb CMCT among the three groups after treatment (all p < 0.05); Compared with baseline, rMT and upper limb CMCT were significantly decreased on the first day and at 30 days after treatment in all three groups (all p < 0.001); Compared with the sham stimulation group, the dTMS group had lower rMT and upper limb CMCT on the first day and at 30 days after treatment (all p < 0.05); Compared with the rTMS group, the dTMS group showed lower rMT and upper limb CMCT on the first day after treatment (p < 0.05), at 30 days after treatment, rMT was lower (p < 0.05), while no significant difference was found in upper limb CMCT (p > 0.05).Conclusion(1) Both high-frequency dTMS and rTMS can improve upper limb motor dysfunction after basal ganglia cerebral infarction to some extent, and the therapeutic effect of dTMS lasts longer; (2) dTMS has a certain rehabilitative effect on lower limb motor and balance function; (3) The mechanisms underlying the improvement of motor dysfunction after basal ganglia cerebral infarction by high-frequency dTMS and rTMS may be associated with increased excitability of the affected cerebral cortex, enhanced function of the corticospinal tract pathway. In addition, dTMS can directly act on deeper and wider brain regions; (4) Both high-frequency dTMS and rTMS are safe.

Acupuncture modulates the microbiota-gut-brain axis to treat irritable bowel syndrome: a mechanistic exploration

Irritable bowel syndrome (IBS) is a common functional gastrointestinal disorder involving dysregulation of the microbiota-gut-brain (MGB) axis. Acupuncture effectively alleviates IBS symptoms, yet its underlying mechanisms remain incompletely understood. This review synthesizes current evidence to propose a mechanistic framework by which acupuncture treats IBS through MGB axis modulation. We systematically examine: (1) MGB axis dysfunction in IBS pathophysiology across neural, endocrine, and immune pathways; (2) acupuncture’s modulation of gut microbiota structure (alpha/beta diversity, specific bacterial genera); (3) functional consequences including enhanced short-chain fatty acid production and tryptophan metabolism; (4) causal evidence from fecal microbiota transplantation; (5) correlations between microbiota changes and clinical improvement. Key findings reveal that acupuncture induces “convergent remodeling” of microbial structure toward a healthy profile, exerts “bidirectional regulation” on beneficial and harmful bacteria, and establishes a “niche selection” mechanism via neuro-immune pathways. These microbiota-mediated effects integrate with neural, endocrine, and immune pathways, forming a “point-to-surface” networked regulatory pattern that explains acupuncture’s dual efficacy in alleviating both gastrointestinal and psychological symptoms. This review provides a novel theoretical framework for understanding acupuncture’s therapeutic mechanisms and supports its clinical application in IBS management.

Without getting under your skin: non-invasive stimulation activates the vagus nerve

Cervical non-invasive vagus nerve stimulation (nVNS) has emerged as a practical neuromodulation approach with FDA-cleared indications in primary headache disorders, yet skepticism persists over whether transcutaneous stimulation can reliably engage vagal fibers or whether observed benefits reflect nonspecific cervical activation. Here, we synthesize converging anatomical, biophysical, physiological, and clinical evidence demonstrating that nVNS does, in fact, activate vagal pathways without surgical implantation. We first review cervical vagus anatomy and the biophysical basis for target engagement, including ultrasound-measured nerve depth and multi-scale computational models showing that clinically relevant stimulation can recruit predominantly large myelinated vagal fibers. We then integrate mechanistic evidence across complementary modalities: functional imaging consistently modulates canonical vagal projection sites (including brainstem nuclei), electrophysiology demonstrates peripheral vagal recruitment and centrally transmitted evoked responses, immune studies reveal reproducible suppression of pro-inflammatory cytokines consistent with cholinergic anti-inflammatory reflex engagement, and autonomic biomarkers show shifts toward increased parasympathetic tone. Finally, we contextualize these mechanistic findings with sham-controlled randomized trials in cluster headache and migraine, where nVNS repeatedly outperforms sham for acute and preventive outcomes with a favorable safety profile. Together, these independent lines of evidence form a coherent mechanistic fingerprint that is difficult to reconcile with placebo or superficial muscle stimulation accounts. We conclude that nVNS provides a credible, scalable means of accessing vagal neurophysiology and represents a clinically validated, paradigm-shifting advance in bioelectronic medicine.

Predicting early neurological deterioration in acute branch atheromatous disease without reperfusion therapy: a machine learning model

BackgroundAcute branch atheromatous disease (BAD) is one of the leading contributors to morbidity and disability in Asia, and early neurological deterioration (END) is common in affected patients. This study aimed to establish machine learning models to predict the risk of END in patients without reperfusion therapy.MethodsPatients with acute BAD who did not receive reperfusion therapy were retrospectively enrolled. Core predictive features were selected by LASSO regression with bootstrap stability assessment, and we used seven machine learning algorithms to build models. XGBoost was selected based on validation performance, nested cross-validation, and 1,000-iteration bootstrap validation. A spline logistic regression model served as the non-linear baseline. SHAP analysis was used to explain the model and develop a simple scoring system. Model discrimination was assessed using the area under the receiver operating characteristic curve (AUC), and clinical utility was evaluated using decision curve analysis (DCA).ResultsA total of 369 patients were included in our research. We screened predictive factors with LASSO regression and ultimately identified five key variables. These included maximum infarct area, lactate dehydrogenase (LDH), number of infarct slices, admission systolic blood pressure (SBP), and neutrophil count. The XGBoost model achieved the best overall performance, with AUC of 0.927 in the training set and 0.846 in the validation set. Nested cross-validation yielded an unbiased AUC of 0.866 (95% CI: 0.817–0.925), and bootstrap validation produced a mean OOB AUC of 0.855 (95% CI: 0.760–0.941). The scoring system stratified patients into low (0–6 points), intermediate (7–13 points), and high (14–20 points) risk groups. DCA demonstrated favorable clinical utility. SHAP analysis also indicated that maximum infarct area and LDH were the top two predictors of END.ConclusionAn XGBoost-based prediction model and a simple scoring system, integrating maximum infarct area, LDH, number of infarct slices, admission SBP, and neutrophil count, provide reliable END risk prediction for acute BAD patients without reperfusion therapy.

A high-dimensional atlas of parvalbumin interneuron soma morphology in mouse visual and somatosensory cortex

Parvalbumin-positive (PV+) inhibitory interneurons are central components of experience-dependent plasticity in the visual cortex (V1). Anatomically, they are distributed across cortical layers and are traditionally classified as basket, chandelier, or bipolar cells based on dendritic and axonal arborization patterns. In parallel, physiological, transcriptomic, connectomic, and multimodal studies have revealed substantial diversity within PV+ populations, raising the question of how this diversity is organized within cortical circuits. In contrast, light microscopy studies based on soma labeling have primarily quantified PV+ cells using size and density, and the potential of soma morphology to capture this diverse organization remains unclear. To address this, we developed a high-throughput, data-driven approach to quantify PV+ soma morphology in > 14,000 cells from mouse V1 and somatosensory cortex (S1). Using 97 morphological features combined with clustering, phenotyping, and laminar mapping, we identified structured diversity in PV+ somas. PV+ cells were organized into 13 morphological clusters along partially independent gradients of size and shape. Phenotyping identified four size and five shape categories that describe PV+ cell diversity across cortical areas. Mapping these categories onto cortical layers revealed a structured organization in which specific morphologies are enriched within distinct laminar compartments. This organization aligns with cortical architecture and suggests that PV+ interneuron morphology is systematically related to circuit structure. These findings demonstrate that substantial morphological information can be extracted from standard PV+ labeling approaches using quantitative analysis. Together, this work provides a high-dimensional atlas of PV+ interneuron soma morphology in mouse V1 and S1 and establishes a framework for linking cellular anatomy to circuit organization and experience-dependent plasticity.

Tracking the longitudinal course of physiologic and mental health functioning among individuals in substance use disorder treatment

IntroductionMental health monitoring is crucial to long-term recovery in substance use disorder (SUD) treatment; however, little is known about how changes in physiological indicators align with changes in self-reported mental health over time.MethodsWe examined longitudinal associations of resting heart rate (RHR) and heart rate variability (HRV) collected via a WHOOP® photoplethysmography device with self-reported stress, anxiety, and depressive symptoms among individuals in SUD treatment. Participants (N = 59) continuously wore the device and completed at least two mental health and stress assessments during the first month of residential treatment. ResultsLinear regression results indicated favorable changes in mental health and/or physiologic metrics, with notable heterogeneity in concurrent subject-level trends. Among participants with decreased RHR (better physiological functioning), 39% (N=23) also endorsed decreased stress, 42% (N=25) decreased anxiety, and 39% (N=23) improved depressive symptoms. Of those with increased HRV (greater stress adaptability), 39% (N=23) endorsed decreased stress, 39% (N=23) improved anxiety, and 41% (N=24) reduced depressive symptoms.DiscussionConcurrent changes in physiologic and mental health metrics during the first month of treatment varied across participants. These findings highlight the importance of integrating subjective mental health measures with physiological indicators to capture clinically relevant change during early SUD treatment.

Longitudinal changes in MMN and P3 during emotional processing in adolescents who engage in NSSI: a 12-week follow-up study

BackgroundAdolescents with nonsuicidal self-injury (NSSI) often show deficits in negative emotion regulation. Within the dual-process framework of implicit and explicit emotion regulation, these deficits may reflect an automatic bias toward negative information and insufficient later-stage controlled regulation. Whether routine clinical intervention modifies both early automatic detection and later controlled evaluation of negative emotional information remains unclear.MethodsIn a longitudinal sample of 32 adolescents with NSSI, we examined changes in mismatch negativity (MMN) and P3 components elicited during an emotional oddball task before and after a 12-week clinical intervention. At baseline and Week 12, participants completed clinical assessments and a two-choice visual emotional oddball task containing neutral standards and negative, positive, and neutral deviants while EEG was recorded. MMN (Fz, 240–300 ms) and P3 (Pz, 450–650 ms) amplitudes were derived from deviant-minus-standard difference waves, indexing pre-attentive deviance detection and controlled evaluation, respectively.ResultsThe intervention selectively modulated neurocognitive processing of negative deviants. Self-injury ideation days (p = .036) and NSSI episode frequency (p = .007) both significantly decreased. MMN absolute amplitude to negative stimuli decreased significantly (p = .047), suggesting attenuated automatic salience detection, whereas P3 amplitude to negative stimuli increased significantly (p = .027), indicating enhanced controlled evaluation. Responses to positive and neutral conditions and behavioural performance remained stable. No baseline ERP–clinical correlations emerged; post-intervention, NSSI ideation days correlated with positive P3 (r = 0.380, p = .032), and episode frequency with neutral P3 (r = 0.461, p = .008).ConclusionsA 12-week intervention may attenuate automatic negative bias while enhancing controlled evaluative processing in adolescents with NSSI, consistent with the implicit–explicit framework. MMN and P3 may serve as candidate biomarkers and temporally sensitive tools for tracking intervention response when behavioural performance is stable.

Case Report: Reintroduction of winged infusion sets to needle syringe programs – advancing equity in harm reduction

BackgroundIn January 2023, following a 20-year ban, the New South Wales (NSW) Needle and Syringe Program (NSP) Guidelines permitted NSPs to provide winged infusion sets (‘butterflies’) and larger-volume syringes (‘barrels’) commonly used for injecting methadone liquid intravenously.ApproachThe Kirketon Road Centre (KRC) integrated the distribution of ‘butterflies and barrels’ (referred to as ‘butterfly packs’) into its service delivery in April 2023 and developed resources in collaboration with the Uniting Medically Supervised Injecting Centre (MSIC) and the NSW Users and AIDS Association (NUAA). This case report presents an audit of service delivery and self-reported client outcomes following the introduction of ‘butterfly packs’.OutcomesSince April 2023, KRC has been consistently supplying ‘barrels’ and ‘butterflies’ to clients who inject methadone. The proportion of NSP visits by clients who reported methadone as the last drug injected increased significantly after the program’s introduction (from 2.2 to 4.0 methadone-related attendances per month), and so did the amount of corresponding injecting equipment and verbal education sessions. Among 18 survey respondents, the majority reported improved health or safety since accessing butterfly packs (89%).ConclusionEquipment for intravenous methadone administration can be safely and efficiently distributed by NSP programs. Access to harm reduction services should be based on vulnerability and need, with services tailored accordingly. The policy reversal allowing ‘butterflies and barrels’ and service innovations such as KRC’s program have addressed a longstanding gap, advancing equity in harm reduction for people who inject methadone in NSW.

Association between plasma proBDNF levels and cognitive impairment in patients with alcohol dependence: a case–control and longitudinal study

BackgroundAlcohol dependence is frequently accompanied by cognitive impairment. Brain-derived neurotrophic factor (BDNF) signaling plays a critical role in synaptic plasticity, while the precursor form, proBDNF, has been increasingly implicated in neurodegenerative and psychiatric disorders. However, the association between plasma proBDNF levels and cognitive impairment in alcohol dependence remains unclear.MethodsEighty male patients with alcohol dependence and forty-two matched healthy controls were enrolled. Plasma proBDNF levels were measured via enzyme-linked immunosorbent assay (ELISA). Cognitive function was assessed using the Mini-Mental State Examination (MMSE), the Modified Wisconsin Card Sorting Test (M-WCST), and the Verbal Fluency Test (VFT). Forty-one patients were reassessed after four weeks of abstinence. Group comparisons and correlation analyses were performed.ResultsPatients with alcohol dependence exhibited significantly elevated plasma proBDNF levels and impaired cognitive performance compared with controls. Plasma proBDNF levels were positively correlated with alcohol consumption severity, and linked to global cognitive deficits alongside nuanced executive performance variations. After four weeks of abstinence, plasma proBDNF levels decreased and cognitive performance improved; however, changes in proBDNF were weakly associated with cognitive recovery.ConclusionsElevated plasma proBDNF levels are associated with alcohol dependence severity and cognitive impairment, suggesting that proBDNF may serve as a peripheral biomarker reflecting the dynamic neurocognitive status in alcohol dependence.