Developer docs
API playgroundTry for free, no card

Search company profiles

Mats Research

Full company profile

uuid02lszm4

Namestring
Mats Research
Legal namestring
MATS Research Inc.
Company typeenum
Private
Founded yearint
2021
Descriptiontext

MATS Research is a 501(c)(3) nonprofit founded in 2021 that operates an AI safety research fellowship headquartered in Berkeley, California, with a second office in London and a regional research management hub in Washington DC. The program runs a 10-12 week intensive fellowship that pairs competitively selected researchers (4-7% acceptance rate) with mentors from leading AI safety organizations including Anthropic, OpenAI, Google DeepMind, Redwood Research, and ARC. Fellows receive stipends of $1,250/week, compute resources of $2,000/week, housing, meals, and travel support, and can apply for a 6-12 month funded extension based in London.

The organization's core product is the fellowship program itself, structured around research streams in mechanistic interpretability, AI control, scalable oversight, evaluations, red-teaming, AI governance, biosecurity, and systems security. MATS has trained 527+ researchers since late 2021, producing 200+ publications with 12,000+ collective citations and an organizational h-index of 50. Three primary "customers" exist: prospective fellows (who receive training rather than pay), research mentors (who participate largely pro bono from frontier labs), and charitable funders (Coefficient Giving, Good Ventures, BERI, Survival and Flourishing Fund). The program generates no commercial revenue and is funded entirely through philanthropic donations and donated compute credits from OpenAI, Anthropic, and Google.

MATS has built a strong field-building flywheel where 80% of alumni remain in AI alignment, transparency, and security roles and 10% have co-founded new organizations such as Apollo Research and Timaeus. Notable alumni placements include research roles at Anthropic, the UK AI Security Institute, and METR, indicating that MATS functions as a primary talent pipeline for the technical AI safety ecosystem. The company is governed by a board including representatives from Future of Life Foundation, Golden Gate Institute for AI, and Longview Philanthropy, and is led by co-founders Ryan Kidd (CEO) and Christian Smith (CFRO).

Short descriptiontext

MATS Research is a 501(c)(3) nonprofit operating an AI safety research fellowship in Berkeley and London. It trains competitively selected researchers under mentorship from frontier AI labs, producing 200+ publications and feeding talent into organizations like Anthropic, UK AISI, and Apollo Research.

Operating statusenum
Operating
Ownership categoryenum
Headcount rangeband
1–10
akta.pro rankint
HeadquartersBerkeley, United States
HQ citystring
Berkeley
HQ countrystring
United States
HQ regionstring
North America
Markets served

Serves global market

Offices3 records

Each record includes

City, Country, Type, Description, Source

Keyword5 values
AI safety fellowship, research mentorship program, alignment research training, nonprofit research fellowship, AI policy education
Industry3 codes
1AI & Machine Learning (Foundations & Applied)
CodeEDAMACAGPrimaryYes
2Safety, Alignment & Content Moderation (Guardrails, Red Teaming)
CodeHDAAACAIPrimaryNo
3Responsible AI, AI Governance & Compliance Services
CodeBPAEAHAJPrimaryNo
NAICS code2 codes
  • Scientific Research and Development Services5417
  • Computer Training611420
SIC code1 code
  • Services-Commercial Physical & Biological Research8731
Product category
AI Safety Research Fellowship
Social media profiles1 record
GTM motion1 record

Each record includes

Type, Description, Source

Revenue model2 records
1Donations and Charitable Grants
TypeData Monetisation
Description

MATS Research is a 501(c)(3) public charity that operates entirely on donations from foundations and individuals. The organization does not generate commercial revenue. Major funding comes from Coefficient Giving, Good Ventures Foundation, Berkeley Existential Risk Initiative (BERI), Survival and Flourishing Fund, Cooperative AI Foundation, Founders Pledge, Foresight Institute, Manifund, and individual donors. Compute credits are also donated by AI companies including OpenAI, Anthropic, and Google.

matsprogram.org
2General Applications
TypeData Monetisation
Description

Open application process for prospective Research Managers and other staff positions

matsprogram.org
Marketing channels6 records

Each record includes

Title, Type, Stage, Description, Source

Distribution channels2 records

Each record includes

Title, Type, Scope, Target buyer, Description, Source

Cost components6 values
Personnel, Operations, Technology or R&D, Marketing or Sales, Infrastructure, Supply Chain
Pricing details2 tiers
1Fellowship with full financial support
ModelOtherBilling cadencePay-as-you-go
Notes

Fellows receive $1,250/week stipend ($12,500 for 10-week program), $2,000/week compute resources, travel to/from Berkeley, housing, and catered lunch/dinner on weekdays. During extension phase, fellows receive housing stipend.

matsprogram.org
2Extension program support
ModelOtherBilling cadenceMulti-year contract
Notes

Extension fellows receive monthly stipend, housing stipend, and compute budget. MATS arranges funding covering all extension phase expenses.

matsprogram.org
GTM typeB2C
B2C
Offering typeServices
Services
Core offering1 text field

MATS Research runs a 10-12 week in-person research fellowship in AI alignment, transparency, and security, hosted at offices in Berkeley, CA and London, UK, with an optional 6-12 month funded Extension Program. Fellows are matched with mentors from leading AI labs and research organizations, receive a weekly stipend (approximately $1,250/week), dedicated compute budget (approximately $2,000/week), travel, housing, and meals, and produce original AI safety research as the core output of the program.

Differentiator
Functional benefit
Problem solved
Quantifiable outcome1 of 7 values shown
  • 527+ researchers trained since late 2021
+6 more records
Product overview1 text field

MATS Research operates as a fellowship program, not a traditional software product company. Its primary offering is the MATS Program, a 10-week intensive research fellowship in AI alignment, transparency, and security, paired with an optional 6-12 month Extension Program for continued research. Fellows receive mentorship from researchers at leading AI organizations (Anthropic, OpenAI, Google DeepMind, Redwood Research, ARC), financial support (stipend and compute budget), workspace, housing, and meals. The program connects emerging talent with the AI safety research ecosystem to produce impactful research.

Product and service2 records
1MATS Program (Main Fellowship)
CategoryAI safety research fellowship
Description

A 10-12 week intensive in-person research fellowship that connects selected researchers with world-class mentors in AI alignment, transparency, and security. Located at MATS offices in Berkeley, CA and London, UK, fellows receive a weekly stipend (~$1,250/week), a compute budget (~$2,000/week), travel, housing, and meals while conducting original AI safety research.

2MATS Extension Program
CategoryAI safety research fellowship extension
Description

A 6-12 month funded research extension for select MATS fellows, providing continued mentorship, monthly stipend, housing stipend, and compute budget. Approximately 80% of eligible fellows are admitted to the extension phase.

Scale indicator11 records

Each record includes

Type, Value, Description, Source

Partnership7 partners
1London Initiative for Safe AI (LISA)
Strategic tierCoreTypeStrategic or Co-development PartnerAnnounced on2024-01-01
Description

Co-founded by MATS CEO Ryan Kidd and CFO Christian Smith. LISA is a UK-based initiative for safe AI research that collaborates with MATS on extending AI safety research capacity and field-building efforts.

matsprogram.org
2BlueDot Impact
Strategic tierSupportingTypeChannel Partner/ Reseller/ DistributorAnnounced on2024-01-01
Description

Educational organization running the Technical AI Safety Course and other programs. Cross-promotes with MATS and is a recommended pathway for prospective MATS applicants.

matsprogram.org
3ARENA
Strategic tierSupportingTypeChannel Partner/ Reseller/ DistributorAnnounced on2024-01-01
Description

Technical bootcamp for AI safety that is recommended for prospective MATS applicants to develop technical ML and software engineering skills.

matsprogram.org
4ML4Good
Strategic tierSupportingTypeChannel Partner/ Reseller/ DistributorAnnounced on2024-01-01
Description

AI safety educational program recommended for applicants to build skills before applying to MATS.

matsprogram.org
5Alignment Research Center (ARC)
Strategic tierCoreTypeStrategic or Co-development PartnerAnnounced on2021-01-01
Description

ARC is a key mentor organization hosting MATS fellows in Berkeley offices. MATS scholars work with ARC on mechanistic explanations of neural network behavior under the Theory track.

matsprogram.org
6Redwood Research
Strategic tierCoreTypeStrategic or Co-development PartnerAnnounced on2021-01-01
Description

Partner organization providing mentorship and research projects for MATS fellows. Alumni including Cody Rushing continue research with Redwood on AI Control.

matsprogram.org
7SPAR
Strategic tierSupportingTypeChannel Partner/ Reseller/ DistributorAnnounced on2021-01-01
Description

AI safety research program that serves as a feeder pathway for MATS. Alumni have conducted research under SPAR mentorship before joining MATS.

matsprogram.org
Recent move6 records

Each record includes

Date, Type, Title, Description, Source

Expansion highlight5 records

Each record includes

Type, Description

Peers10 records
1BlueDot Impact
TypeDirect peer
Description

Runs the Technical AI Safety Course and other AI safety educational programs that serve as recommended feeders for MATS applicants. Directly comparable mission of training early-career AI safety talent and overlapping applicant pools.

2ARENA (Alignment Research Engineer Accelerator)
TypeDirect peer
Description

Technical AI safety bootcamp recommended as a pre-MATS pathway. Operates a similar selective, intensive training model for aspiring AI safety researchers with comparable curriculum focus.

3AI Safety Camp
TypeDirect peer
Description

Runs team-based AI safety research programs with mentorship from field practitioners. Directly comparable fellowship-style training model targeting the same early-career AI safety researcher segment.

4SPAR
TypeDirect peer
Description

Supervised Program for Alignment Research serving as a feeder pathway to MATS. Operates supervised AI safety research projects for aspiring researchers with comparable mentorship-driven structure.

5ML4Good
TypeDirect peer
Description

AI safety educational program recommended for prospective MATS applicants. Comparable mission of building pre-fellowship technical ML and AI safety skills in early-career researchers.

6Redwood Research
TypeDirect peer
Description

AI control research organization that hosts MATS fellows and provides mentorship. Directly comparable technical AI alignment research focus and overlapping research talent pool.

7Alignment Research Center (ARC)
TypeDirect peer
Description

Core mentor organization hosting MATS fellows in Berkeley on mechanistic interpretability. Directly comparable AI alignment research mission with deep operational integration with MATS.

8Apollo Research
TypeEmerging player
Description

London-based technical alignment organization founded by MATS alumnus Marius Hobbhahn. Compares as an alumni spin-out conducting frontier AI safety evaluation research with overlapping mission.

9FAR AI
TypeBroad incumbent
Description

AI safety research organization conducting alignment, evaluation, and robustness research. Comparable AI safety research focus with overlapping mentor and alumni relationships (FAR AI board member Josh Jacobson also serves on MATS's board).

10Anthropic
TypeBroad incumbent
Description

Frontier AI lab providing mentors to MATS and hiring significant numbers of MATS alumni. Indirect comparable as the primary destination employer for MATS graduates and a key compute partner.

Market position
Strengths5 records

Each record includes

Headline, Details, Source

Weaknesses4 records

Each record includes

Headline, Details, Source

Competitive moat4 records

Each record includes

Type, Details

Key risks6 records

Each record includes

Headline, Details, Source

Key highlights7 records

Each record includes

Headline, Details, Source

Customer concentration

Classification, Details

Named customers10 records

Each record includes

Name, Industry, Type, Use case, Source, UUID

Segment7 records

Each record includes

Title, Type, Primary, Description, Pain point addressed, Use case, Source

Ideal customer profile4 records

Each record includes

Profile, Firmographic size, Sales motion, Sales cycle length, Buying structure, Purchase trigger, Buyer persona, Geography, Industry vertical, Primary use case, Description, Pain points, Evidence proof points, Target buyer

Technology focused
No
API detail
Has APIbool
No

Docs URL, Description

AI maturity
App detail

Has app

Feature2 records

Each record includes

Title, Differentiator, Description, Source

Core technology
Revenue estimate
Valuation estimate
Number of profiles
Profiles11 records

Each record includes

Name, Designation, Designation category, Overview, Profile commentary, Source

No data
No data
Funding overview

Funding stage, Last funding date, Total funding USD

Funding rounds

Each record includes

Round, Amount USD, Date, Pre money valuation, Total investors, Investors, News

Investors

Each record includes

Name, Type, Date of entry, Rounds participated, Website

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Each record includes

Name, Acquisition type, Announced date, Completed date, Status, Website, News

Investment

Each record includes

Name, Round, Announced date, Lead investor, Website, News

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Mats Research

AI Safety Research Fellowshipmatsprogram.org

MATS Research is a 501(c)(3) nonprofit operating an AI safety research fellowship in Berkeley and London. It trains competitively selected researchers under mentorship from frontier AI labs, producing 200+ publications and feeding talent into organizations like Anthropic, UK AISI, and Apollo Research.

What Mats Research does

MATS Research is a 501(c)(3) nonprofit founded in 2021 that operates an AI safety research fellowship headquartered in Berkeley, California, with a second office in London and a regional research management hub in Washington DC. The program runs a 10-12 week intensive fellowship that pairs competitively selected researchers (4-7% acceptance rate) with mentors from leading AI safety organizations including Anthropic, OpenAI, Google DeepMind, Redwood Research, and ARC. Fellows receive stipends of $1,250/week, compute resources of $2,000/week, housing, meals, and travel support, and can apply for a 6-12 month funded extension based in London.

The organization's core product is the fellowship program itself, structured around research streams in mechanistic interpretability, AI control, scalable oversight, evaluations, red-teaming, AI governance, biosecurity, and systems security. MATS has trained 527+ researchers since late 2021, producing 200+ publications with 12,000+ collective citations and an organizational h-index of 50. Three primary "customers" exist: prospective fellows (who receive training rather than pay), research mentors (who participate largely pro bono from frontier labs), and charitable funders (Coefficient Giving, Good Ventures, BERI, Survival and Flourishing Fund). The program generates no commercial revenue and is funded entirely through philanthropic donations and donated compute credits from OpenAI, Anthropic, and Google.

MATS has built a strong field-building flywheel where 80% of alumni remain in AI alignment, transparency, and security roles and 10% have co-founded new organizations such as Apollo Research and Timaeus. Notable alumni placements include research roles at Anthropic, the UK AI Security Institute, and METR, indicating that MATS functions as a primary talent pipeline for the technical AI safety ecosystem. The company is governed by a board including representatives from Future of Life Foundation, Golden Gate Institute for AI, and Longview Philanthropy, and is led by co-founders Ryan Kidd (CEO) and Christian Smith (CFRO).

Mats Research firmographics

Firmographics
Name
Mats Research
Legal name
MATS Research Inc.
Website
https://www.matsprogram.org
Company type
Private
Founded year
2021
Operating status
Operating
Headcount range
1–10 employees
Short description
MATS Research is a 501(c)(3) nonprofit operating an AI safety research fellowship in Berkeley and London. It trains competitively selected researchers under mentorship from frontier AI labs, producing 200+ publications and feeding talent into organizations like Anthropic, UK AISI, and Apollo Research.
Ownership category
akta.pro rank

Mats Research industry classification

Industry
Product category
AI Safety Research Fellowship
NAICS
Scientific Research and Development Services (5417), Computer Training (611420)
SIC
Services-Commercial Physical & Biological Research (8731)
akta.pro primary industry
AI & Machine Learning (Foundations & Applied) (EDAMACAG)
akta.pro secondary industries
Safety, Alignment & Content Moderation (Guardrails, Red Teaming) (HDAAACAI), Responsible AI, AI Governance & Compliance Services (BPAEAHAJ)

Keywords

  • AI safety fellowship
  • Research mentorship program
  • Alignment research training
  • Nonprofit research fellowship
  • AI policy education

Where Mats Research is headquartered

Location

Headquarters

HQ city
Berkeley
HQ country
United States
HQ region
North America

Offices3 records

Markets served

Mats Research business model

Business model
GTM type
B2C
Offering type
Services
Cost components
Personnel, Operations, Technology or R&D, Marketing or Sales, Infrastructure, Supply Chain

Revenue model

  1. Donations and Charitable Grants: MATS Research is a 501(c)(3) public charity that operates entirely on donations from foundations and individuals. The organization does not generate commercial revenue. Major funding comes from Coefficient Giving, Good Ventures Foundation, Berkeley Existential Risk Initiative (BERI), Survival and Flourishing Fund, Cooperative AI Foundation, Founders Pledge, Foresight Institute, Manifund, and individual donors. Compute credits are also donated by AI companies including OpenAI, Anthropic, and Google.
  2. General Applications: Open application process for prospective Research Managers and other staff positions

Pricing tiers

ModelBillingPrice
OtherPay-as-you-goFellowship with full financial support
OtherMulti-year contractExtension program support

Go-to-market motion1 record

Distribution channels2 records

Marketing channels6 records

Mats Research product offering

Product offering

Core offering

MATS Research runs a 10-12 week in-person research fellowship in AI alignment, transparency, and security, hosted at offices in Berkeley, CA and London, UK, with an optional 6-12 month funded Extension Program. Fellows are matched with mentors from leading AI labs and research organizations, receive a weekly stipend (approximately $1,250/week), dedicated compute budget (approximately $2,000/week), travel, housing, and meals, and produce original AI safety research as the core output of the program.

Product overview

MATS Research operates as a fellowship program, not a traditional software product company. Its primary offering is the MATS Program, a 10-week intensive research fellowship in AI alignment, transparency, and security, paired with an optional 6-12 month Extension Program for continued research. Fellows receive mentorship from researchers at leading AI organizations (Anthropic, OpenAI, Google DeepMind, Redwood Research, ARC), financial support (stipend and compute budget), workspace, housing, and meals. The program connects emerging talent with the AI safety research ecosystem to produce impactful research.

Differentiator

Problem solved

Functional benefit

Products and services

  • MATS Program (Main Fellowship) A 10-12 week intensive in-person research fellowship that connects selected researchers with world-class mentors in AI alignment, transparency, and security. Located at MATS offices in Berkeley, CA and London, UK, fellows receive a weekly stipend (~$1,250/week), a compute budget (~$2,000/week), travel, housing, and meals while conducting original AI safety research.
  • MATS Extension Program A 6-12 month funded research extension for select MATS fellows, providing continued mentorship, monthly stipend, housing stipend, and compute budget. Approximately 80% of eligible fellows are admitted to the extension phase.

Quantifiable outcome

  • 527+ researchers trained since late 2021
  • +6 more outcomes

Companies that use Mats Research

Customer profile

Named customers10 records

Segments7 records

Ideal customer profiles4 records

Mats Research technology and API

Technology

Technology focussed No

API detail

Has API
No
API docs
API detail

Core technology

AI maturity

App detail

Feature2 records

Mats Research partnerships and signals

Strategic signal

Partnerships

Seven partnerships are on record, tiered core and supporting.

  • London Initiative for Safe AI (LISA)coreStrategic or Co-development Partner · 1 January 2024Co-founded by MATS CEO Ryan Kidd and CFO Christian Smith. LISA is a UK-based initiative for safe AI research that collaborates with MATS on extending AI safety research capacity and field-building efforts.
  • BlueDot ImpactsupportingChannel Partner/ Reseller/ Distributor · 1 January 2024Educational organization running the Technical AI Safety Course and other programs. Cross-promotes with MATS and is a recommended pathway for prospective MATS applicants.
  • ARENAsupportingChannel Partner/ Reseller/ Distributor · 1 January 2024Technical bootcamp for AI safety that is recommended for prospective MATS applicants to develop technical ML and software engineering skills.
  • ML4GoodsupportingChannel Partner/ Reseller/ Distributor · 1 January 2024AI safety educational program recommended for applicants to build skills before applying to MATS.
  • Alignment Research Center (ARC)coreStrategic or Co-development Partner · 1 January 2021ARC is a key mentor organization hosting MATS fellows in Berkeley offices. MATS scholars work with ARC on mechanistic explanations of neural network behavior under the Theory track.
  • Redwood ResearchcoreStrategic or Co-development Partner · 1 January 2021Partner organization providing mentorship and research projects for MATS fellows. Alumni including Cody Rushing continue research with Redwood on AI Control.
  • SPARsupportingChannel Partner/ Reseller/ Distributor · 1 January 2021AI safety research program that serves as a feeder pathway for MATS. Alumni have conducted research under SPAR mentorship before joining MATS.

Scale indicators11 records

Recent moves6 records

Expansion highlights5 records

Mats Research competitors and assessment

Company assessment

Direct peers

  • BlueDot Impact: Runs the Technical AI Safety Course and other AI safety educational programs that serve as recommended feeders for MATS applicants. Directly comparable mission of training early-career AI safety talent and overlapping applicant pools.
  • ARENA (Alignment Research Engineer Accelerator): Technical AI safety bootcamp recommended as a pre-MATS pathway. Operates a similar selective, intensive training model for aspiring AI safety researchers with comparable curriculum focus.
  • AI Safety Camp: Runs team-based AI safety research programs with mentorship from field practitioners. Directly comparable fellowship-style training model targeting the same early-career AI safety researcher segment.
  • SPAR: Supervised Program for Alignment Research serving as a feeder pathway to MATS. Operates supervised AI safety research projects for aspiring researchers with comparable mentorship-driven structure.
  • ML4Good: AI safety educational program recommended for prospective MATS applicants. Comparable mission of building pre-fellowship technical ML and AI safety skills in early-career researchers.
  • Redwood Research: AI control research organization that hosts MATS fellows and provides mentorship. Directly comparable technical AI alignment research focus and overlapping research talent pool.
  • Alignment Research Center (ARC): Core mentor organization hosting MATS fellows in Berkeley on mechanistic interpretability. Directly comparable AI alignment research mission with deep operational integration with MATS.

Emerging players

  • Apollo Research: London-based technical alignment organization founded by MATS alumnus Marius Hobbhahn. Compares as an alumni spin-out conducting frontier AI safety evaluation research with overlapping mission.

Broad incumbents

  • FAR AI: AI safety research organization conducting alignment, evaluation, and robustness research. Comparable AI safety research focus with overlapping mentor and alumni relationships (FAR AI board member Josh Jacobson also serves on MATS's board).
  • Anthropic: Frontier AI lab providing mentors to MATS and hiring significant numbers of MATS alumni. Indirect comparable as the primary destination employer for MATS graduates and a key compute partner.

Market position

Strengths5 records

Weaknesses4 records

Competitive moat4 records

Key risks6 records

Key highlights7 records

Customer concentration

Mats Research social profiles

Digital presence

Mats Research financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Mats Research leadership team

Management profile

Number of profiles

Profiles11 records

Mats Research funding detail

Funding detail

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

Mats Research M&A and investment

M&A and investment

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Mats Research

What does Mats Research do?

MATS Research runs a 10-12 week in-person research fellowship in AI alignment, transparency, and security, hosted at offices in Berkeley, CA and London, UK, with an optional 6-12 month funded Extension Program. Fellows are matched with mentors from leading AI labs and research organizations, receive a weekly stipend (approximately $1,250/week), dedicated compute budget (approximately $2,000/week), travel, housing, and meals, and produce original AI safety research as the core output of the program.

Is Mats Research a public or private company?

Mats Research is a private company. It is classified as nonprofit foundation owned and is currently operating.

When was Mats Research founded?

Mats Research was founded in 2021. It employs 1 to 10 people.

Where is Mats Research based?

Mats Research is headquartered in Berkeley, United States, in the North America region.

How does Mats Research make money?

Two revenue lines are on record. Donations and Charitable Grants are the primary driver. The others are general Applications.

Who are Mats Research's main competitors?

Direct peers on record are BlueDot Impact, ARENA (Alignment Research Engineer Accelerator), AI Safety Camp, SPAR, ML4Good, Redwood Research and Alignment Research Center (ARC). Apollo Research is listed as an emerging player. Broad incumbents are FAR AI and Anthropic.

Does Mats Research have an API?

No public API is recorded for Mats Research.

What industry is Mats Research in?

Mats Research's product category is AI Safety Research Fellowship. Its primary akta.pro industry code is EDAMACAG, AI & Machine Learning (Foundations & Applied), with a secondary code of HDAAACAI, Safety, Alignment & Content Moderation (Guardrails, Red Teaming). Its NAICS code is 5417 and its SIC code is 8731.

Unlock the full company data

50 free credits on sign-up, no credit card required.

Contact sales
Live signals
CryptopolitanResearchers use LLMs to link anonymous forum accounts to real names for $1 to $4 per targetResearchers from ETH Zurich, MATS, and Anthropic developed an LLM pipeline that linked 226 of 338 pseudonymous Hacker News users to their LinkedIn profiles at 90% precision. The method cost under $2,000 total, or $1 to $4 per profile, and the authors withheld code and identities.NateAI can now find real names through posts and web searches... A crisis of online anonymity: Nate NewsA study by ETH Zurich, MATS, and Anthropic found AI can de-anonymize pseudonymous accounts using public posts and web searches. In experiments, AI identified 226 of 338 Hacker News users, achieving 67% recall and 90% precision. The researchers warn this lowers cost barriers to identifying scattered personal information.DecryptAI Can Now Doxx Your Anonymous Accounts? Here's What’s Going OnResearchers from ETH Zurich, MATS, and Anthropic published a February paper showing AI can deanonymize pseudonymous accounts, matching Hacker News users to LinkedIn profiles at 90% precision. The attack costs $1–4 per target and uses only web search and reasoning models, without hacking. The authors withheld their code and real identities.The Indian ExpressCan an AI model’s ‘reasoning’ be extracted? New research fuels US-China distillation rowResearchers from the University of Tubingen, Max Planck Institute, MATS Research, and Snyk published a paper revealing a method to extract hidden reasoning traces from frontier AI models. The findings suggest that Chinese developers may have utilized this distillation technique to train open-weight models like Moonshot AI's Kimi K3.