Microsoft Looks Beyond OpenAI, Student Cracks Century-Old Math Problem, and Research Shows How Aspirin Fights Cancer

We're experimenting and would love to hear from you!

In this episode of 'Discover Daily', we begin with a look at Microsoft. The company is strategically diversifying its AI approach beyond its OpenAI partnership by developing its own AI models, particularly the MAI family. These models have performed nearly as well as leading models from OpenAI and Anthropic on commonly accepted benchmarks and are being considered as potential replacements for OpenAI's technology in Microsoft's Copilot AI assistants. Under the leadership of Mustafa Suleyman, who joined Microsoft last year as head of its AI division, the company is also exploring collaborations with other AI companies like xAI, Meta, and DeepSeek to reduce reliance on OpenAI and maintain a competitive edge in the rapidly evolving AI landscape.

Meanwhile, a Penn State University graduate student Divya Tyagi has refined a century-old mathematical problem in wind turbine aerodynamics, potentially advancing wind energy science and turbine design efficiency. Her groundbreaking research, published in Wind Energy Science, presents an amendment to Hermann Glauert's optimum rotor disk solution from the 1920s, offering a more comprehensive mathematical model that provides a realistic depiction of turbine dynamics. Even a 1% increase in power coefficient for large wind turbines could translate to substantial gains in energy output, potentially powering entire communities and leading to more efficient turbine designs for existing wind farms.

Lastly, in a new study from the University of Cambridge published in Nature, researchers have discovered how aspirin enhances the immune system's ability to target and combat metastatic cancer cells. The research reveals that aspirin decreases thromboxane A2 production, which normally suppresses T cells' ability to identify and eliminate cancer cells. By reducing thromboxane A2 levels, aspirin effectively releases T cells from this suppression, enabling them to recognize and attack cancer cells before they establish metastatic tumors. This finding was demonstrated in a mouse model of melanoma, where aspirin-treated mice exhibited significantly fewer metastases compared to control subjects.

From Perplexity's Discover Feed:

https://www.perplexity.ai/page/microsoft-looks-beyond-openai-xcvmz0JyTiipf0BCk8BWgA

https://www.perplexity.ai/page/student-cracks-century-old-mat-wN0I7t44Q8qaalwydsOpEw

https://www.perplexity.ai/page/research-shows-how-aspirin-fig-arCADzAVTfOMpVcKvaD5Vg

**Introducing Perplexity Deep R

Perplexity is the fastest and most powerful way to search the web. Perplexity crawls the web and curates the most relevant and up-to-date sources (from academic papers to Reddit threads) to create the perfect response to any question or topic you’re interested in.

Take the world's knowledge with you anywhere. Available on iOS and Android

Join our growing Discord community for the latest updates and exclusive content.

Follow us on:

• Instagram
• Threads
• X (Twitter)
• YouTube
• 
  Today we're exploring howaspirin might be a powerful tool
  in the fight against cancer.
  But first let's look at whatelse is happening across the
  tech and science landscape.
  Microsoft is making moves todiversify its AI strategy beyond
  
  (00:23):
  its close partnership withOpenAI.
  The tech giant is developingits own AI models, particularly
  the MAI family, which have shownpromising results in recent
  tests.
  These models are beingconsidered as potential
  replacements for OpenAI'stechnology.
  In Microsoft's co-pilot AIassistance, microsoft's MAI
  models have performed nearly aswell as leading models from
  
  (00:46):
  OpenAI and Anthropic on commonlyaccepted benchmarks.
  The company is also creatingits own AI reasoning models
  comparable to OpenAI's, designedto handle complex queries and
  display human-likeproblem-solving abilities.
  This shift in strategy comesunder the leadership of Mustafa
  Suleiman, who joined Microsoftlast year as the head of its AI
  
  (01:06):
  division.
  Suleiman, co-founder ofDeepMind and former CEO of
  Inflection AI, has beeninstrumental in pushing for
  greater independence from OpenAI.
  Microsoft is not just focusingon in-house development.
  The company is also exploringcollaborations with other AI
  companies like XAI, meta andDeepSeek.
  This diversification aims toreduce reliance on OpenAI and
  
  (01:30):
  maintain a competitive edge inthe rapidly evolving AI
  landscape.
  While Microsoft remains a majorinvestor in OpenAI, this
  strategic move signals a shiftin the dynamics between the two
  companies.
  It raises questions about thefuture of their collaboration
  and the potential impact onexisting products that rely on
  OpenAI's technology.
  
  (01:51):
  Moving on to our second story, agraduate student at Penn State
  University has refined acentury-old mathematical problem
  in wind turbine aerodynamics,potentially advancing wind
  energy science and turbinedesign efficiency.
  Divya Tyagi's groundbreakingresearch, published in Wind
  Energy Science, presents anamendment to Hermann Glauert's
  
  (02:11):
  optimum rotor disk solution fromthe 1920s.
  Tyagi's work offers analternative mathematical
  approach, using calculus ofvariations, and provides
  analytical solutions for thrustand bending moment coefficients.
  In other words, tyagi's newmethod gives clear formulas for
  measuring how forces affectstructures, making it easier to
  understand than previousapproaches.
  
  (02:32):
  This comprehensive mathematicalmodel provides a more realistic
  depiction of turbine dynamics.
  It expands on Glauert'soriginal focus by incorporating
  total force and momentcoefficients acting on the rotor
  , as well as how turbine bladesbend under wind pressure.
  Tiagi's research introduces anovel approach to solve for
  
  (02:52):
  optimal flow conditions,recovering Glauert's original
  distributions for axial andangular induction factors.
  Her work also reveals finite,non-zero values for thrust and
  bending moment coefficients astip speed ratio approaches zero.
  More simply, tiagi found abetter way to figure out how
  wind should flow throughturbines, while also showing
  
  (03:14):
  that even when turbine bladesare barely moving, they still
  experience forces that can pushand bend them.
  The practical implications ofthis refinement are significant.
  Even a 1% increase in powercoefficient for large wind
  turbines could translate tosubstantial gains in energy
  output, potentially poweringentire communities.
  This advancement could lead tomore efficient turbine designs
  
  (03:36):
  and increased energy productionfrom existing wind farms.
  Now let's dive into our mainstory of the day.
  A significant study from theUniversity of Cambridge
  published in Nature hasuncovered how aspirin could play
  
  (03:58):
  a role in preventing cancerfrom spreading.
  The research reveals thataspirin enhances the immune
  system's ability to target andcombat metastatic cells.
  Aspirin enhances the immunesystem's ability to target and
  combat metastatic cells.
  This breakthrough not onlyhighlights aspirin's potential
  as a cancer-fighting medication,but also paves the way for
  
  (04:20):
  creating more precise therapiesaimed at halting cancer
  metastasis.
  Aspirin's cancer-fightingpotential stems from its ability
  to decrease thromboxane A2production, a clotting factor
  produced by platelets in thebloodstream.
  Thromboxane A2 normallyactivates a protein called
  ARHGEF1 in T cells, suppressingtheir ability to identify and
  
  (04:40):
  eliminate metastatic cancercells.
  By reducing thromboxane A2levels, aspirin effectively
  releases T cells from thissuppression, enabling them to
  recognize and attack cancercells before they establish
  metastatic tumors.
  Dr Jay Yang, one of theresearchers, described this
  discovery as a eureka moment.
  To demonstrate aspirin'seffectiveness, researchers used
  
  (05:03):
  a mouse model of melanoma.
  Mice treated with aspirinexhibited significantly fewer
  metastases compared to controlsubjects.
  This reduction in cancer spreadwas directly linked to
  aspirin's ability to preventthromboxane A2 from impairing
  T-cell function.
  Professor Rahul Roychudhuri,who led the study, emphasized a
  
  (05:24):
  critical therapeutic window ofopportunity when cancer cells
  are particularly vulnerable toimmune attack.
  This finding could be crucialin developing targeted therapies
  to prevent cancer metastasis.
  Building on this research, theADASPRIN clinical trial aims to
  determine if aspirin can stop ordelay the recurrence of
  early-stage cancers.
  This trial follows previouslarge randomized controlled
  
  (05:46):
  studies that unexpectedly foundlower cancer mortality rates
  among daily aspirin users.
  But aspirin can cause seriousside effects in some individuals
  , including bleeding and stomachulcers.
  Researchers emphasize theimportance of consulting a
  doctor before starting anyaspirin regimen.
  Looking ahead, this discoveryopens up exciting possibilities
  
  (06:08):
  for future cancer treatments.
  Researchers are exploring thepotential of aspirin or other
  drugs targeting the thromboxaneA2 pathway as more affordable
  alternatives to antibody-basedtherapies.
  As metastasis is responsiblefor approximately 90% of
  cancer-related deaths worldwide,this breakthrough could lead to
  significant advancements incancer prevention and treatment
  
  (06:30):
  strategies, particularly inaddressing the critical early
  stages of metastasis.
  That's all for today's episodeof Discover Daily.
  Before we conclude, a briefnote on our evolution at
  
  (06:50):
  Perplexity.
  Our new deep research feature,launched earlier this month, now
  analyzes hundreds of sources inminutes.
  Think of it as deploying apersonal research team through
  our web and mobile platforms.
  This cutting-edge tool combinesautonomous reasoning with rapid
  processing to deliverexhaustive reports on
  specialized topics.
  Deep research excels at expertlevel tasks across various
  
  (07:11):
  domains, from finance andmarketing to product research,
  and is available on our desktopand mobile apps.
  Thank you for listening and foryour fan mail.
  We'll be back with more storiesthat shape our world.
  Until then, stay curious.