Stephen Greer
@stepheng29.bsky.social
Industry Consultant at SAS Institute. Interested in the tech industry, AI/ML, data science, fintech. 49ers fan.
Nobody can agree on a definition for AI Agents
No one knows what the hell an AI agent is | TechCrunch
AI agents are all the rage. But no one knows exactly what an agent is, partly because companies define them radically differently.
techcrunch.com
March 18, 2025 at 9:58 PM
Nobody can agree on a definition for AI Agents
There's a lot more to browsers than just the default search engine. As a customer acquisition strategy this might make sense, but what else would be here that would make it better than simply going to the web UI?
Perplexity teases a web browser called Comet | TechCrunch
AI-powered search engine Perplexity says it's building its own web browser. It'll be called Comet, and it's set to arrive 'soon.'
techcrunch.com
March 4, 2025 at 4:00 PM
There's a lot more to browsers than just the default search engine. As a customer acquisition strategy this might make sense, but what else would be here that would make it better than simply going to the web UI?
This is a great long list of resources to explore agent models...
GitHub - francedot/acu: A curated list of resources about AI agents for Computer Use, including research papers, projects, frameworks, and tools.
A curated list of resources about AI agents for Computer Use, including research papers, projects, frameworks, and tools. - francedot/acu
github.com
February 11, 2025 at 4:22 PM
This is a great long list of resources to explore agent models...
Google released a list of 321 real-world use cases for Generative AI. Definitely worth a read through.
Real-world gen AI use cases from the world's leading organizations | Google Cloud Blog
Gen AI is everywhere, as top companies, governments, researchers, and startups showcase how they're already using Google's AI solutions to enhance their work.
cloud.google.com
January 29, 2025 at 12:42 PM
Google released a list of 321 real-world use cases for Generative AI. Definitely worth a read through.
Reposted by Stephen Greer
Exploring the gap between what LLMs really know vs what people think they know
www.nature.com/articles/s42...
www.nature.com/articles/s42...
What large language models know and what people think they know - Nature Machine Intelligence
Understanding how people perceive and interpret uncertainty from large language models (LLMs) is crucial, as users often overestimate LLM accuracy, especially with default explanations. Steyvers et al...
www.nature.com
January 21, 2025 at 4:08 PM
Exploring the gap between what LLMs really know vs what people think they know
www.nature.com/articles/s42...
www.nature.com/articles/s42...
Singapore telco offering Perplexity Pro as a bundle, just like free Spotify or Netflix.
www.techinasia.com/news/singtel...
www.techinasia.com/news/singtel...
Tech in Asia - Connecting Asia's startup ecosystem
www.techinasia.com
January 21, 2025 at 4:17 PM
Singapore telco offering Perplexity Pro as a bundle, just like free Spotify or Netflix.
www.techinasia.com/news/singtel...
www.techinasia.com/news/singtel...
Michael Barr's resignation as fed VC of banking supervision is worrying to many why watch the industry.
This may be the end of independent bank regulation and may mean a turnover of Fed VC with each administration.
This may be the end of independent bank regulation and may mean a turnover of Fed VC with each administration.
January 9, 2025 at 2:39 PM
Michael Barr's resignation as fed VC of banking supervision is worrying to many why watch the industry.
This may be the end of independent bank regulation and may mean a turnover of Fed VC with each administration.
This may be the end of independent bank regulation and may mean a turnover of Fed VC with each administration.
This is very cool, a guide to building an OS in 1000 lines.
Intro | OS in 1,000 Lines
Write your first operating system from scratch, in 1K LoC.
operating-system-in-1000-lines.vercel.app
January 9, 2025 at 2:28 PM
This is very cool, a guide to building an OS in 1000 lines.
OpenAI o3 models achieve 87.5% on the ARC-AGI Eval, absolutely blowing away any of the previous models.
December 22, 2024 at 12:47 AM
OpenAI o3 models achieve 87.5% on the ARC-AGI Eval, absolutely blowing away any of the previous models.
This presentation from @fromedome.bsky.social is always a good one every year
(Love a 40-year chart!)
From my latest Consumer Trends Highlights newsletter: newconsumer.com/2024/11/shop...
From my latest Consumer Trends Highlights newsletter: newconsumer.com/2024/11/shop...
Shopping never ends
Consumer spending is looking good heading into the holiday rush, and online grocery is growing again. The latest Consumer Trends Highlights.
newconsumer.com
December 17, 2024 at 8:04 PM
This presentation from @fromedome.bsky.social is always a good one every year
I'm very curious how real these claims actually are and tend to think this is PR for a future IPO against recent valuation adjustments. If a company could actually carve out 20% of their workforce with AI then they'd be crazy not to spin it off into a what could easily be a unicorn.
Klarna's CEO says it stopped hiring thanks to AI but still advertises many open positions | TechCrunch
Klarna CEO Sebastian Siemiatkowski recently told Bloomberg TV that his company essentially stopped hiring a year ago and credited generative AI for
techcrunch.com
December 17, 2024 at 6:41 PM
I'm very curious how real these claims actually are and tend to think this is PR for a future IPO against recent valuation adjustments. If a company could actually carve out 20% of their workforce with AI then they'd be crazy not to spin it off into a what could easily be a unicorn.
YC's request for startups Winter 2025
If you were to start an economy today from scratch....
If you were to start an economy today from scratch....
December 10, 2024 at 5:00 PM
YC's request for startups Winter 2025
If you were to start an economy today from scratch....
If you were to start an economy today from scratch....
Reposted by Stephen Greer
That input image was the ultimate robustness test for computer vision and ML in 1989. Imagine trying some structural pattern recognition on this, which was popular at that time. But, @yann-lecun.bsky.social's convnets solved it 😍
He posted the video on LinkedIn:
www.youtube.com/watch?v=H0oE...
December 5, 2024 at 11:33 AM
That input image was the ultimate robustness test for computer vision and ML in 1989. Imagine trying some structural pattern recognition on this, which was popular at that time. But, @yann-lecun.bsky.social's convnets solved it 😍
He posted the video on LinkedIn:
www.youtube.com/watch?v=H0oE...
Anthropic released the Model Context Protocol (MCP), which intended to bridge the gap between data sources/systems and LLMs. The full protocol is open-source and already comes with integrations for things like Google Drive, Slack, Postgres, Git and more
Introducing the Model Context Protocol
The Model Context Protocol (MCP) is an open standard for connecting AI assistants to the systems where data lives, including content repositories, business tools, and development environments. Its aim...
www.anthropic.com
December 4, 2024 at 3:57 PM
Anthropic released the Model Context Protocol (MCP), which intended to bridge the gap between data sources/systems and LLMs. The full protocol is open-source and already comes with integrations for things like Google Drive, Slack, Postgres, Git and more
Reposted by Stephen Greer
How awesome is Bluesky?
The tools around it are going to become so amazing:
Check out this list~ repost this if you know someone who might find it useful.
🔵🌍
github.com/fishttp/awes...
The tools around it are going to become so amazing:
Check out this list~ repost this if you know someone who might find it useful.
🔵🌍
github.com/fishttp/awes...
GitHub - fishttp/awesome-bluesky: A list of all known tools available for the Bluesky platform
A list of all known tools available for the Bluesky platform - fishttp/awesome-bluesky
github.com
December 3, 2024 at 5:24 PM
How awesome is Bluesky?
The tools around it are going to become so amazing:
Check out this list~ repost this if you know someone who might find it useful.
🔵🌍
github.com/fishttp/awes...
The tools around it are going to become so amazing:
Check out this list~ repost this if you know someone who might find it useful.
🔵🌍
github.com/fishttp/awes...
The whole "debanking" debate is based on a fundamental misunderstanding of the role of the regulators and risk management at a bank. Risk management is difficult and expensive; most FIs error on the side of caution for unfamiliar risk
This is a great overview and explanation below from Jason Mikula
This is a great overview and explanation below from Jason Mikula
Synapse Investor Marc Andreessen Says CFPB "Terrorizes" Fintech, Crypto To Protect Big Banks
a16z's Sprawling Fintech & Crypto Investments Stand To Benefit From Lighter Regulation
fintechbusinessweekly.substack.com
December 3, 2024 at 5:19 PM
The whole "debanking" debate is based on a fundamental misunderstanding of the role of the regulators and risk management at a bank. Risk management is difficult and expensive; most FIs error on the side of caution for unfamiliar risk
This is a great overview and explanation below from Jason Mikula
This is a great overview and explanation below from Jason Mikula
Apparently LLMs perform much better in complex tasks when they are told to imitate rather than purely reason. 🎭
🔑 Acting-based prompting:
Treat LLMs as performers, prompts as scripts.
🔑 Acting-based prompting:
Treat LLMs as performers, prompts as scripts.
December 3, 2024 at 1:41 PM
Apparently LLMs perform much better in complex tasks when they are told to imitate rather than purely reason. 🎭
🔑 Acting-based prompting:
Treat LLMs as performers, prompts as scripts.
🔑 Acting-based prompting:
Treat LLMs as performers, prompts as scripts.
Reposted by Stephen Greer
As I've complained before in this here platform, AI gen'd images are taking over websites that us artists use for references. Here's an example, screenshot courtesy of a reddit post (from user MetaKnowing). Same thing is happening on DeviantArt, Pinterest etc. with no clear way to filter them out.
October 7, 2024 at 10:37 PM
As I've complained before in this here platform, AI gen'd images are taking over websites that us artists use for references. Here's an example, screenshot courtesy of a reddit post (from user MetaKnowing). Same thing is happening on DeviantArt, Pinterest etc. with no clear way to filter them out.
AI Agents Stack, Nov 2024
November 30, 2024 at 3:45 PM
AI Agents Stack, Nov 2024
Perplexity keeps shipping very interesting features
November 26, 2024 at 4:52 PM
Perplexity keeps shipping very interesting features
Reposted by Stephen Greer
A quick tip on coding with AI.
If you do not get the code you are looking for by the third turn (i.e., request/response, clarification/response, clarification/response), start a new chat session. It does not get any better.
If you do not get the code you are looking for by the third turn (i.e., request/response, clarification/response, clarification/response), start a new chat session. It does not get any better.
November 25, 2024 at 6:24 AM
A quick tip on coding with AI.
If you do not get the code you are looking for by the third turn (i.e., request/response, clarification/response, clarification/response), start a new chat session. It does not get any better.
If you do not get the code you are looking for by the third turn (i.e., request/response, clarification/response, clarification/response), start a new chat session. It does not get any better.
The consensus with Apple Intelligence seems to be that it's average at best. Especially the weirdness of notification summaries. This is still early days and no doubt will improve--I'm excited to see improvements to Siri or rumored cross-app programmable agentic behavior
Apple Intelligence notification summaries are honestly pretty bad
Summaries are often wrong, usually odd, sometimes funny, rarely helpful.
arstechnica.com
November 25, 2024 at 2:11 PM
The consensus with Apple Intelligence seems to be that it's average at best. Especially the weirdness of notification summaries. This is still early days and no doubt will improve--I'm excited to see improvements to Siri or rumored cross-app programmable agentic behavior
Reposted by Stephen Greer
While Bluesky looks like other social apps on the surface, it’s actually quite different. It’s an open network.
I blogged about it here:
I blogged about it here:
Benefits of an open network
emilyliu.me
November 24, 2024 at 11:53 PM
While Bluesky looks like other social apps on the surface, it’s actually quite different. It’s an open network.
I blogged about it here:
I blogged about it here:
DynaSaur: The AI agent that evolves by coding
→ Writes new Python functions when needed
→ Reuses & optimizes actions via embeddings
→ Builds a growing library of tools
Results:
🏆 Tops GAIA leaderboard
📈 +81.59% w/ human + AI tools
🚀 Excels in complex tasks (level 2 and 3)
→ Writes new Python functions when needed
→ Reuses & optimizes actions via embeddings
→ Builds a growing library of tools
Results:
🏆 Tops GAIA leaderboard
📈 +81.59% w/ human + AI tools
🚀 Excels in complex tasks (level 2 and 3)
November 25, 2024 at 1:50 PM
DynaSaur: The AI agent that evolves by coding
→ Writes new Python functions when needed
→ Reuses & optimizes actions via embeddings
→ Builds a growing library of tools
Results:
🏆 Tops GAIA leaderboard
📈 +81.59% w/ human + AI tools
🚀 Excels in complex tasks (level 2 and 3)
→ Writes new Python functions when needed
→ Reuses & optimizes actions via embeddings
→ Builds a growing library of tools
Results:
🏆 Tops GAIA leaderboard
📈 +81.59% w/ human + AI tools
🚀 Excels in complex tasks (level 2 and 3)
Drag & Drop prompt engineering IDE to eval LLMs for agentic workflow.
All OS and No-code.
All OS and No-code.
November 21, 2024 at 5:27 PM
Drag & Drop prompt engineering IDE to eval LLMs for agentic workflow.
All OS and No-code.
All OS and No-code.