Andy Tseng
@andytseng.bsky.social
2.4K followers 730 following 1.8K posts
Director of IT and Digital Strategy at @whiteheadinstitute.bsky.social, MIT. PhD in CS. Turning data into discoveries. Fan of practical tech that works. Tea drinker, research enabler, innovation chaser. Views are my own. He/him.
Posts Media Videos Starter Packs
Pinned
andytseng.bsky.social
Hey everyone! Quick intro: I’m passionate about #AI, #data, and transforming research into real-world impact. I specialise in research IT infrastructure, #HPC, and data-driven projects, worked in the UK, France, Australia and the US. If you're excited by emerging tech and research, let’s connect!
andytseng.bsky.social
Australia can do this - a future powered by 100% renewables is possible. Phasing out coal isn’t easy, but investment, grid upgrades, and bold policy make it real. As an Australian, I’m hopeful and all in for a cleaner future.
#Renewables #Australia #CleanEnergy #ClimateAction #NetZero
Australia’s March Toward 100 Percent Clean Energy
The country’s grid operator says shifting from coal to clean power is not only possible but inevitable. The work there could provide a road map for other countries.
www.wired.com
andytseng.bsky.social
No way!!! 👏🏻👏🏻👏🏻
andytseng.bsky.social
I'm just waiting for someone in Godzilla costume!
andytseng.bsky.social
What’s next, I wonder...?
andytseng.bsky.social
Perspective matters! @ourworldindata.org's report reveals how #MainstreamMedia amplifies rare tragedies while ignoring the real causes of death affecting millions. Fear sells, truth saves. Research and question what you see. #MediaBias

🔗 ourworldindata.org/does-the-news-reflect-what-we-die-from
Infographic by "Our World in Data", comparing U.S. causes of death in 2023 with media coverage by The New York Times, The Washington Post, and Fox News. Shows heart disease and cancer as leading causes but underreported, while terrorism and homicide are heavily covered despite low death rates.
andytseng.bsky.social
This whole "orbital sunlight" thing is insane. Who asked for light pollution on demand? It's a fundamentally terrible idea that trashes the night sky - just stop! 🤦🏻‍♂️
#LightPollution #SaveTheNight
A US startup plans to deliver ‘sunlight on demand’ after dark. Can it work – and would we want it to?
Satellites beaming sunlight down to Earth sound like science fiction – and they have astronomers very worried.
theconversation.com
andytseng.bsky.social
Meta pushing AI to "go 5x faster" isn’t strategy, it’s desperation. Speed without direction = 5x the bugs, tech debt & burnout. If AI is really that magic, maybe it can explain why the #metaverse flopped - or better yet, fix it!
#Meta #AI #TechEthics #Productivity
Meta Tells Its Metaverse Workers to Use AI to ‘Go 5X Faster’
Mark Zuckerberg's metaverse chief is urging employees to adopt AI across every workflow as part of a broader shift inside the company.
www.wired.com
andytseng.bsky.social
Nicely laid out by @ianlmorgan.bsky.social: NIH leadership is failing early-career researchers - overburdened, overlooked, and underfunded. If we want innovations, we must invest in the people driving them.
#NIH #USPol #ScienceMatters #ResearchSky #AcademicSky #Accountability
ianlmorgan.bsky.social
My op-ed just came out discussing how NIH leadership is failing early career researchers and, importantly, how we can push back and take our future into our own hands.
NIH Leadership Is Failing Early Career Researchers
Halted training programs, funding chaos, and hiring freezes jeopardize advancement options for early career researchers.
www.the-scientist.com
andytseng.bsky.social
Techbros lack the sophistication and intelligence to understand and connect with women - that’s the issue!
andytseng.bsky.social
Great to see MIT stand up for principle - defending merit, free academic expression, and research independence. Funding should be merit-based, not political. Proud to be part of this community.
#USPol #MIT #SallyKornbluth #AcademicFreedom #ResearchIntegrity #HigherEd #AcademicSky #ResearchSky
Regarding the Compact | MIT Organization Chart
Dear members of the MIT community,
orgchart.mit.edu
andytseng.bsky.social
So true! I increasingly feel that way every time I see major announcements about AI infrastructure.
andytseng.bsky.social
Let’s remember the importance of checking in with ourselves and those around us. Be kind, speak openly, and know that seeking support is a sign of strength. You are not alone! 💚
#WorldMentalHealthDay #WMHD2025 #MentalHealthMatters #YouAreNotAlone
Green ribbon with "WFMH" and sunburst symbol next to "World Mental Health Day" text, promoting global mental health awareness.
andytseng.bsky.social
Not just you, my friends, there seem to be new initiatives every day. However, I’m concerned about the environmental impact - energy, water, etc. - especially considering the proposed size and number of the data centers these companies are committed to building.
andytseng.bsky.social
Seriously, the pace of AI infrastructure is bonkers. Altman confirms even *after* the #Stargate/#Oracle/#Nvidia/#AMD partnerships, OpenAI has more massive deals coming. They're basically securing all the compute capacity in the galaxy.
#OpenAI #AGI #AIWars
Even after Stargate, Oracle, Nvidia, and AMD, OpenAI has more big deals coming soon, Sam Altman says | TechCrunch
By some estimates, OpenAI has this year inked $1 trillion worth of infrastructure deals. Altman says to expect even more.
techcrunch.com
andytseng.bsky.social
Just 250 poisoned docs can backdoor a 13B-parameter LLM and no need to control a percentage of the training data! Big models? Small models? Same risk. Time to rethink our defenses. Great work, @anthropic.com, @theturing.bsky.social and UK AI Security Institute.
#AISecurity #LLM #Cybersecurity
A small number of samples can poison LLMs of any size
Anthropic research on data-poisoning attacks in large language models
www.anthropic.com
andytseng.bsky.social
Another bold AI claim from #Zendesk - “AI agent can solve 80% of support issues”? What counts as “solved”? Redirecting to a how-to article isn’t real resolution. Where’s the real evidence and human backup when things go wrong?
#AI #AIHype #Skepticism
Zendesk says its new AI agent can solve 80% of support issues | TechCrunch
Zendesk has launched a new autonomous support agent that should solve 80% of support issues without human intervention.
techcrunch.com
andytseng.bsky.social
C'mon #ANU, you can and should do better - sharp insight from @richarddenniss.bsky.social.
#HigherEd #Australia #UniversityCrisis #AcademicSky
australiainstitute.org.au
New Australia Institute analysis reveals there is no financial crisis at the Australian National University.

Audited accounts show the ANU made a $90 million surplus in 2024 and increased the value of its net assets.

#auspol @richarddenniss.bsky.social
andytseng.bsky.social
I just tried it now, and it seems to be working for me.
A screenshot of the results page from a Bluesky search query of 'from:@btravern.bsky.social'.
andytseng.bsky.social
(10/08 11:58 EDT) - #Bluesky search is up and running again, panic over!
andytseng.bsky.social
(10/08 09:20 EDT) - #Bluesky search is currently returning no skeets at all. Really hope this is just a coincidence and nothing more.
#SearchFail
andytseng.bsky.social
In case you’re wondering why #Blacksky is trending, @nicomaramckay.com’s piece offers deep insight. I don’t have all the facts, but what I’ve seen is worrying. Silence from @jay.bsky.team & @bsky.app only adds confusion. I support #ATProtocol & #Decentralization, but #Bluesky users deserve better.
nicomaramckay.com
tl;dr: Bluesky began with genuine promise: a new social media platform where users could feel safer, but inconsistent moderation quickly eroded trust.

When systemically marginalized users pushed back, Bluesky’s leadership minimized concerns and began to target them for bans.
andytseng.bsky.social
In case you’re wondering why #Blacksky is trending, @nicomaramckay.com’s piece offers deep insight. I don’t have all the facts, but what I’ve seen is worrying. Silence from @jay.bsky.team & @bsky.app only adds confusion. I support #ATProtocol & #Decentralization, but #Bluesky users deserve better.
nicomaramckay.com
tl;dr: Bluesky began with genuine promise: a new social media platform where users could feel safer, but inconsistent moderation quickly eroded trust.

When systemically marginalized users pushed back, Bluesky’s leadership minimized concerns and began to target them for bans.
Bluesky’s CEO meltdown: How leadership continues to fail its most marginalized users
tl;dr: Bluesky began with genuine promise: a new social media platform where systemically marginalized users could feel safer, but…
plutopsyche.medium.com