Jason Murphy
@jason-murphy.bsky.social
Interested in AI ethics, tools, and machine learning
Sometimes seen in CODE Magazine doing silly stuff with AI.
Sometimes seen in CODE Magazine doing silly stuff with AI.
A true, privacy-forward phone?
Sign me up, friend.
Sign me up, friend.
Privacy stalwart Nicholas Merrill spent a decade fighting an FBI surveillance order. Now he wants to sell you phone service—without knowing almost anything about you.
A New Anonymous Phone Carrier Lets You Sign Up With Nothing but a Zip Code
Privacy stalwart Nicholas Merrill spent a decade fighting an FBI surveillance order. Now he wants to sell you phone service—without knowing almost anything about you.
wrd.cm
December 4, 2025 at 5:49 PM
A true, privacy-forward phone?
Sign me up, friend.
Sign me up, friend.
The threat to the internet's underwater backbone is no secret, but the author provides a set of solutions that should be enacted *yesterday*.
The Underwater Cables That Carry the Internet Are in Trouble
If you ask the average American, “where does the internet come from?” the answer you would most likely get would be from space, via satellites. Wrong. The vast majority of information that flows acros...
www.bloomberg.com
December 1, 2025 at 3:17 PM
The threat to the internet's underwater backbone is no secret, but the author provides a set of solutions that should be enacted *yesterday*.
Not to be outdone by Google ...
Anthropic unveils Claude Opus 4.5, its latest AI model following $350 billion valuation
Anthropic says its new Claude Opus 4.5 excels at coding, using computers and assisting users with complex enterprise tasks.
www.cnbc.com
November 24, 2025 at 8:14 PM
Not to be outdone by Google ...
Are the big players - OpenAI, Anthropic, Meta - still chasing scaling as the path to AGI?
Why Cohere's ex-AI research lead is betting against the scaling race | TechCrunch
Cohere's former VP of AI research, Sara Hooker, is launching a new startup to build AI models that can adapt to their environment.
techcrunch.com
October 24, 2025 at 4:52 PM
Are the big players - OpenAI, Anthropic, Meta - still chasing scaling as the path to AGI?
This will probably be outpaced/semi-obsolete within 9 months.
Do I want one anyway? Yes. Absolutely.
Do I want one anyway? Yes. Absolutely.
Nvidia sells tiny new computer that puts big AI on your desktop
The 1 petaflop DGX Spark system runs AI models with 200 billion parameters locally for $4K.
arstechnica.com
October 14, 2025 at 8:37 PM
This will probably be outpaced/semi-obsolete within 9 months.
Do I want one anyway? Yes. Absolutely.
Do I want one anyway? Yes. Absolutely.
"The consumer does not care if what they’re looking at is AI. We’ve got to embrace it, but ethically and legally."
He's right. There's a *very* vocal contingent of Anti-AI folks online, but Sora quickly became the #1 app on the Apple Store. That's the real market.
He's right. There's a *very* vocal contingent of Anti-AI folks online, but Sora quickly became the #1 app on the Apple Store. That's the real market.
‘Blumhouse Is Not for Sale’: Jason Blum on 15 Years of Horror Hits, Losing ‘Texas Chainsaw,’ Reinventing ‘Saw’ and More
Jason Blum breaks down Blumhouse's 15-year history and what's next for the iconic horror studio in a new Variety cover story.
variety.com
October 8, 2025 at 11:25 PM
"The consumer does not care if what they’re looking at is AI. We’ve got to embrace it, but ethically and legally."
He's right. There's a *very* vocal contingent of Anti-AI folks online, but Sora quickly became the #1 app on the Apple Store. That's the real market.
He's right. There's a *very* vocal contingent of Anti-AI folks online, but Sora quickly became the #1 app on the Apple Store. That's the real market.
Can't wait to play around with OpenAI's Agent Builder.
Every iteration of agents gets more intuitive. Very exciting!
Every iteration of agents gets more intuitive. Very exciting!
OpenAI Platform
Explore developer resources, tutorials, API docs, and dynamic examples to get the most out of OpenAI's platform.
platform.openai.com
October 7, 2025 at 3:20 PM
Can't wait to play around with OpenAI's Agent Builder.
Every iteration of agents gets more intuitive. Very exciting!
Every iteration of agents gets more intuitive. Very exciting!
And I *just* finished writing an article about MCP where I mention that threat actors like this are inevitable.
First Malicious MCP in the Wild: The Postmark Backdoor That's Stealing Your Emails | Koi Blog
www.koi.security
September 27, 2025 at 3:56 PM
And I *just* finished writing an article about MCP where I mention that threat actors like this are inevitable.
Fascinating article about the parallels between biology and code.
Is Life a Form of Computation?
Alan Turing and John von Neumann saw it early: the logic of life and the logic of code may be one and the same.
thereader.mitpress.mit.edu
September 25, 2025 at 2:42 PM
Fascinating article about the parallels between biology and code.
Establishing tech barriers between nations in this race is counter-productive.
And predictable.
And predictable.
What China Banning Nvidia Chips Means for the AI Race - Decrypt
After years of U.S. sanctions, China moves to ban Nvidia overnight, betting Huawei and homegrown chips are enough to win the AI race.
decrypt.co
September 18, 2025 at 3:15 PM
Establishing tech barriers between nations in this race is counter-productive.
And predictable.
And predictable.
I'd love to see a non-monetized, "walled garden" approach to the web. If we're going to silo everything, let's create silos that truly benefit the people within them.
Google admits the open web is in ‘rapid decline’
One Google exec recently said the web is “thriving.”
www.theverge.com
September 11, 2025 at 2:20 PM
I'd love to see a non-monetized, "walled garden" approach to the web. If we're going to silo everything, let's create silos that truly benefit the people within them.
I'm sure he has the appropriate skills, history, and certs to excel at the job, right?
... right?
... right?
New cyber director Cairncross calls on industry to help put 'America First' in cyberspace #cybersecurity #hacking #news #infosec #security #technology #privacy
New cyber director Cairncross calls on industry to help put 'America First' in cyberspace
“We've admired the problem for too long, and now it's time to do something about it,” National Cyber Director Sean Cairncross said about the cybersecurity threat environment.
therecord.media
September 10, 2025 at 3:19 PM
I'm sure he has the appropriate skills, history, and certs to excel at the job, right?
... right?
... right?
"AlterEgo is a non-invasive, wearable, peripheral neural interface that allows humans to converse in natural language with machines, artificial intelligence assistants, services, and other people without any voice ... simply by articulating words internally."
Project Overview ‹ AlterEgo – MIT Media Lab
AlterEgo is a non-invasive, wearable, peripheral neural interface that allows humans to converse in natural language with machines, artificial intel…
www.media.mit.edu
September 10, 2025 at 3:06 PM
"AlterEgo is a non-invasive, wearable, peripheral neural interface that allows humans to converse in natural language with machines, artificial intelligence assistants, services, and other people without any voice ... simply by articulating words internally."
Much of the hand-wringing about AI is filthy with ableism.
"For people with dyslexia in any setting, AI assistants might serve as writing aids that go beyond traditional spell-checkers, potentially helping with sentence structure and organizing thoughts without requiring specialized software."
"For people with dyslexia in any setting, AI assistants might serve as writing aids that go beyond traditional spell-checkers, potentially helping with sentence structure and organizing thoughts without requiring specialized software."
Why accessibility might be AI’s biggest breakthrough
UK study findings may challenge assumptions about who benefits most from AI tools.
arstechnica.com
September 9, 2025 at 5:21 PM
Much of the hand-wringing about AI is filthy with ableism.
"For people with dyslexia in any setting, AI assistants might serve as writing aids that go beyond traditional spell-checkers, potentially helping with sentence structure and organizing thoughts without requiring specialized software."
"For people with dyslexia in any setting, AI assistants might serve as writing aids that go beyond traditional spell-checkers, potentially helping with sentence structure and organizing thoughts without requiring specialized software."
This was a fun and nuanced conversation. It's always nice to hear a real discussion about Gen AI without all of the online vitriol that has flooded the agora.
This was such a fun conversation!
www.youtube.com/watch?v=X2M9...
My main message: The question of whether or not we should use Generative AI shouldn't be framed simply through the lens of harm-benefit analysis of the tech, but grounded in today's broader sociopolitical context.
(contd.)
www.youtube.com/watch?v=X2M9...
My main message: The question of whether or not we should use Generative AI shouldn't be framed simply through the lens of harm-benefit analysis of the tech, but grounded in today's broader sociopolitical context.
(contd.)
September 8, 2025 at 1:49 PM
This was a fun and nuanced conversation. It's always nice to hear a real discussion about Gen AI without all of the online vitriol that has flooded the agora.
I'm an avowed protopianist.
But the road to get there is going to be long and dark.
But the road to get there is going to be long and dark.
‘Godfather of AI’ says the technology will create massive unemployment and send profits soaring — ‘that is the capitalist system’
"We are at a point in history where something amazing is happening, and it may be amazingly good, and it may be amazingly bad."
www.yahoo.com
September 8, 2025 at 1:38 PM
I'm an avowed protopianist.
But the road to get there is going to be long and dark.
But the road to get there is going to be long and dark.
Can we please get more shrugs from LLMs?
LLMs “hallucinate” because training rewards confident guesses.
Teach models to prioritize humility, not obsequiousness.
LLMs “hallucinate” because training rewards confident guesses.
Teach models to prioritize humility, not obsequiousness.
Why language models hallucinate
OpenAI’s new research explains why language models hallucinate. The findings show how improved evaluations can enhance AI reliability, honesty, and safety.
openai.com
September 7, 2025 at 10:25 PM
Can we please get more shrugs from LLMs?
LLMs “hallucinate” because training rewards confident guesses.
Teach models to prioritize humility, not obsequiousness.
LLMs “hallucinate” because training rewards confident guesses.
Teach models to prioritize humility, not obsequiousness.
I attended a panel recently where a few of the participants railed against Generative AI. They presented reasoned, thoughtful arguments, but at the core, these arguments weren't really about AI.
They were about capitalism.
#ai #genai #aiethics
They were about capitalism.
#ai #genai #aiethics
'Godfather of AI' says the technology will create massive unemployment and send profits soaring — 'that is the capitalist system'
"We are at a point in history where something amazing is happening, and it may be amazingly good, and it may be amazingly bad."
fortune.com
September 7, 2025 at 1:27 PM
Many AI companies are already branching out into robotics and have been for some time.
Is robotics the future? I've heard a lot about widespread adoption in the near future. How likely is that?
Is robotics the future? I've heard a lot about widespread adoption in the near future. How likely is that?
Why Runway is eyeing the robotics industry for future revenue growth | TechCrunch
Runway is building up a robotics-focused team and fine-tuning its existing models for robotics and self-driving car customers.
techcrunch.com
September 4, 2025 at 5:18 PM
Many AI companies are already branching out into robotics and have been for some time.
Is robotics the future? I've heard a lot about widespread adoption in the near future. How likely is that?
Is robotics the future? I've heard a lot about widespread adoption in the near future. How likely is that?
My new article for CODE Magazine is live! I run some rudimentary experiments with Veo3.
Veo 3: The Clockwork Hand
Jason explores the transformative impact of generative AI on video production, specifically through Google's Veo 3 model. Beginning with the crude outputs of early generative AI tools, he charts th...
www.codemag.com
September 2, 2025 at 3:50 PM
My new article for CODE Magazine is live! I run some rudimentary experiments with Veo3.
Once again, the takeaway here is that our leadership lacks the will to do anything about it.
The AI Summit Where Everyone Agreed on Bad News
Exclusive: At a summit attended by staff from OpenAI and DeepMind, the economic risks posed by AI came into focus.
time.com
September 1, 2025 at 1:55 PM
Once again, the takeaway here is that our leadership lacks the will to do anything about it.
How exactly did their 'chief ethicist' approve this?
I can only imagine any ethicists at Meta get locked out of the room, mocked, and given a swirly every day.
I can only imagine any ethicists at Meta get locked out of the room, mocked, and given a swirly every day.
Meta’s AI rules have let bots hold ‘sensual’ chats with children
An internal Meta policy document reveals the social-media giant’s rules for chatbots, which have permitted provocative behavior on topics including sex and race.
www.reuters.com
August 27, 2025 at 2:08 PM
How exactly did their 'chief ethicist' approve this?
I can only imagine any ethicists at Meta get locked out of the room, mocked, and given a swirly every day.
I can only imagine any ethicists at Meta get locked out of the room, mocked, and given a swirly every day.
Fantastic!
Now give it away.
Now give it away.
Google DeepMind's Weather Lab, launched in June, showed superior accuracy in forecasting Hurricane Erin's path up to 72 hours ahead, beating traditional models (Eric Berger/Ars Technica)
Main Link | Techmeme Permalink
Main Link | Techmeme Permalink
August 26, 2025 at 6:42 PM
Fantastic!
Now give it away.
Now give it away.
I just learned about this monthly webinar. Particularly interested in AI & Africa and hoping to learn more about decolonial data sets.
Are you ready?
1 Day | 4 Virtual Events | 8 AI Experts | Aug 26, 2025
Trustworthy AI: From Principles to Procurement Practices
AI & Africa: Digital Security - Auditing the Machine
To Use Generative AI or Not to Use it?
AI & Current Events | Available on Substack
open.substack.com/pub/womenina...
1 Day | 4 Virtual Events | 8 AI Experts | Aug 26, 2025
Trustworthy AI: From Principles to Procurement Practices
AI & Africa: Digital Security - Auditing the Machine
To Use Generative AI or Not to Use it?
AI & Current Events | Available on Substack
open.substack.com/pub/womenina...
1 Day | 4 Virtual Events | 8 AI Experts
On August 26, WAIE & WAIE+ bring together experts to discuss AI Ethics topics that matter to 99% of humanity
open.substack.com
August 26, 2025 at 2:07 PM
I just learned about this monthly webinar. Particularly interested in AI & Africa and hoping to learn more about decolonial data sets.