Sherveen
@sherveen.bsky.social
The squeeze is worth it, even if there's no juice.
Founder @ Free Agency.
Build your AI fluency → http://aimuscle.com
For those who notice nuance → http://noticethenuance.com
YouTube (tech, AI, venture, society) → http://youtube.com/@sherveenshow
Founder @ Free Agency.
Build your AI fluency → http://aimuscle.com
For those who notice nuance → http://noticethenuance.com
YouTube (tech, AI, venture, society) → http://youtube.com/@sherveenshow
Tech Twitter is kind of turning on Marc Andreessen tonight.
There may be hope yet, everyone.
There may be hope yet, everyone.
November 9, 2025 at 4:29 AM
Tech Twitter is kind of turning on Marc Andreessen tonight.
There may be hope yet, everyone.
There may be hope yet, everyone.
The dude who receives personalized healthcare from the manosphere oligarchs has something to say.
Adjacency is complacency. If you support this admin, you are an enabler.
Adjacency is complacency. If you support this admin, you are an enabler.
November 5, 2025 at 3:27 PM
The dude who receives personalized healthcare from the manosphere oligarchs has something to say.
Adjacency is complacency. If you support this admin, you are an enabler.
Adjacency is complacency. If you support this admin, you are an enabler.
It should bother us, not encourage us, that people like Chamath and Andreessen were talking about this, can talk about this, and will continue to talk about this...
... while contributing to its acceleration unemotionally.
... while contributing to its acceleration unemotionally.
November 5, 2025 at 2:26 PM
It should bother us, not encourage us, that people like Chamath and Andreessen were talking about this, can talk about this, and will continue to talk about this...
... while contributing to its acceleration unemotionally.
... while contributing to its acceleration unemotionally.
Nothing can stand in the way
of the power of millions of voices
calling for change
We have been told we cannot do this
by a chorus of cynics
In the unlikely story that is America,
there has never been anything false about hope
Fired up,
ready to go
Our time is now!
of the power of millions of voices
calling for change
We have been told we cannot do this
by a chorus of cynics
In the unlikely story that is America,
there has never been anything false about hope
Fired up,
ready to go
Our time is now!
November 5, 2025 at 2:41 AM
Nothing can stand in the way
of the power of millions of voices
calling for change
We have been told we cannot do this
by a chorus of cynics
In the unlikely story that is America,
there has never been anything false about hope
Fired up,
ready to go
Our time is now!
of the power of millions of voices
calling for change
We have been told we cannot do this
by a chorus of cynics
In the unlikely story that is America,
there has never been anything false about hope
Fired up,
ready to go
Our time is now!
You know what, he's not wrong.
November 4, 2025 at 8:49 PM
You know what, he's not wrong.
Specialized, oligarch-directed healthcare for the few.
For the rest? Thoughts, prayers, and an ACA repeal.
For the rest? Thoughts, prayers, and an ACA repeal.
November 3, 2025 at 1:03 PM
Specialized, oligarch-directed healthcare for the few.
For the rest? Thoughts, prayers, and an ACA repeal.
For the rest? Thoughts, prayers, and an ACA repeal.
SWE-1.5 is way better + way faster than Composer 1.
The multi-agent interface + plan-to-build flow is way better than Cascade.
Cursor and Windsurf are both still super compelling products trading worthy jabs at each other.
The multi-agent interface + plan-to-build flow is way better than Cascade.
Cursor and Windsurf are both still super compelling products trading worthy jabs at each other.
November 3, 2025 at 6:33 AM
SWE-1.5 is way better + way faster than Composer 1.
The multi-agent interface + plan-to-build flow is way better than Cascade.
Cursor and Windsurf are both still super compelling products trading worthy jabs at each other.
The multi-agent interface + plan-to-build flow is way better than Cascade.
Cursor and Windsurf are both still super compelling products trading worthy jabs at each other.
I asked the smartest AI models to project the results of the NYC mayoral. The results...
October 30, 2025 at 12:36 AM
I asked the smartest AI models to project the results of the NYC mayoral. The results...
tmux is all you need.
October 28, 2025 at 12:34 PM
tmux is all you need.
Oh no, someone's building a product that attempts to be more emotionally aware and meet some people where they might need or want to be met.
What horrors for the free market to let humans reach for tools of support, should they want it.
What horrors for the free market to let humans reach for tools of support, should they want it.
October 28, 2025 at 1:17 AM
Oh no, someone's building a product that attempts to be more emotionally aware and meet some people where they might need or want to be met.
What horrors for the free market to let humans reach for tools of support, should they want it.
What horrors for the free market to let humans reach for tools of support, should they want it.
Odyssey-2: instantly steerable, interactive AI video.
This is incredible, and everything's about to change. Ya gotta try it: experience.odyssey.ml
This is incredible, and everything's about to change. Ya gotta try it: experience.odyssey.ml
October 27, 2025 at 5:57 PM
Odyssey-2: instantly steerable, interactive AI video.
This is incredible, and everything's about to change. Ya gotta try it: experience.odyssey.ml
This is incredible, and everything's about to change. Ya gotta try it: experience.odyssey.ml
On a mission to make sure "AI agent" means somethin'.
October 25, 2025 at 1:18 PM
On a mission to make sure "AI agent" means somethin'.
No. I've had to hate-watch this shit for like, 7 years. I'm invested now. It's an infected part of my identity. You can't take it away.
Mike Majlak is Wolf Blitzer, Joe Rogan is Larry King, PBD is Bill O'Reilly, Tim Dillon is Stephen Colbert, Destiny is Jon Stewart, and this is CNN.
Mike Majlak is Wolf Blitzer, Joe Rogan is Larry King, PBD is Bill O'Reilly, Tim Dillon is Stephen Colbert, Destiny is Jon Stewart, and this is CNN.
October 23, 2025 at 7:48 AM
No. I've had to hate-watch this shit for like, 7 years. I'm invested now. It's an infected part of my identity. You can't take it away.
Mike Majlak is Wolf Blitzer, Joe Rogan is Larry King, PBD is Bill O'Reilly, Tim Dillon is Stephen Colbert, Destiny is Jon Stewart, and this is CNN.
Mike Majlak is Wolf Blitzer, Joe Rogan is Larry King, PBD is Bill O'Reilly, Tim Dillon is Stephen Colbert, Destiny is Jon Stewart, and this is CNN.
Cuomo is Rabois is Horowitz is Gurley is Sacks.
Zohran is Alexis is Graham is Sacca is no one.
It will continue this way. The last node will remain at imbalance by nature of the third node. I won't be elaborating further.
Zohran is Alexis is Graham is Sacca is no one.
It will continue this way. The last node will remain at imbalance by nature of the third node. I won't be elaborating further.
October 23, 2025 at 7:31 AM
Cuomo is Rabois is Horowitz is Gurley is Sacks.
Zohran is Alexis is Graham is Sacca is no one.
It will continue this way. The last node will remain at imbalance by nature of the third node. I won't be elaborating further.
Zohran is Alexis is Graham is Sacca is no one.
It will continue this way. The last node will remain at imbalance by nature of the third node. I won't be elaborating further.
I watched Cuomo on Impaulsive.
I watched Mamdani on Flagrant.
Game is done.
I watched Mamdani on Flagrant.
Game is done.
October 23, 2025 at 7:16 AM
I watched Cuomo on Impaulsive.
I watched Mamdani on Flagrant.
Game is done.
I watched Mamdani on Flagrant.
Game is done.
For those looking to better understand AI agents --
It's important to recognize the base LLM (ex. GPT-5) at the center of any agentic product (ex. Claude Code or ChatGPT Atlas).
Outcome quality can be driven by the strength of the model, the 'scaffold,' or the combination.
It's important to recognize the base LLM (ex. GPT-5) at the center of any agentic product (ex. Claude Code or ChatGPT Atlas).
Outcome quality can be driven by the strength of the model, the 'scaffold,' or the combination.
October 22, 2025 at 2:08 PM
For those looking to better understand AI agents --
It's important to recognize the base LLM (ex. GPT-5) at the center of any agentic product (ex. Claude Code or ChatGPT Atlas).
Outcome quality can be driven by the strength of the model, the 'scaffold,' or the combination.
It's important to recognize the base LLM (ex. GPT-5) at the center of any agentic product (ex. Claude Code or ChatGPT Atlas).
Outcome quality can be driven by the strength of the model, the 'scaffold,' or the combination.
I really don't understand why Google can't serve Deep Think reliably to users paying $250/month.
Not only is there a tight quota, but over the past few weeks, I get this message about half the time.
Not only is there a tight quota, but over the past few weeks, I get this message about half the time.
October 22, 2025 at 12:21 PM
I really don't understand why Google can't serve Deep Think reliably to users paying $250/month.
Not only is there a tight quota, but over the past few weeks, I get this message about half the time.
Not only is there a tight quota, but over the past few weeks, I get this message about half the time.
Realizing... as my usage of AI became heavy w/ GPT-5 Pro, I stopped thanking the models as much as I used to.
Ya don't want to send it off for 10m overthinking your gratitude!
Ya don't want to send it off for 10m overthinking your gratitude!
October 22, 2025 at 3:31 AM
Realizing... as my usage of AI became heavy w/ GPT-5 Pro, I stopped thanking the models as much as I used to.
Ya don't want to send it off for 10m overthinking your gratitude!
Ya don't want to send it off for 10m overthinking your gratitude!
This. You just have to take setup seriously, but once you do, Codex is meaningfully capable of self-iterative debugging.
October 18, 2025 at 5:47 PM
This. You just have to take setup seriously, but once you do, Codex is meaningfully capable of self-iterative debugging.
Other than subagents being able to call on skills (presumably), it's unclear to me how skills differentiate from subagents in Claude Code.
If Anthropic wants expanded use by non-technical users, there should be clearer bifurcation and best practice to its functionality.
If Anthropic wants expanded use by non-technical users, there should be clearer bifurcation and best practice to its functionality.
October 17, 2025 at 11:55 AM
Other than subagents being able to call on skills (presumably), it's unclear to me how skills differentiate from subagents in Claude Code.
If Anthropic wants expanded use by non-technical users, there should be clearer bifurcation and best practice to its functionality.
If Anthropic wants expanded use by non-technical users, there should be clearer bifurcation and best practice to its functionality.
Super interesting.
New ad real estate, persistent in the AI agent chat window and targeted based on 'conversation' context, to enable Amp Free -- $0 version of an otherwise premium product.
Once again, Sourcegraph leading the way on big pricing bets in the age of AI.
New ad real estate, persistent in the AI agent chat window and targeted based on 'conversation' context, to enable Amp Free -- $0 version of an otherwise premium product.
Once again, Sourcegraph leading the way on big pricing bets in the age of AI.
October 15, 2025 at 6:29 PM
Super interesting.
New ad real estate, persistent in the AI agent chat window and targeted based on 'conversation' context, to enable Amp Free -- $0 version of an otherwise premium product.
Once again, Sourcegraph leading the way on big pricing bets in the age of AI.
New ad real estate, persistent in the AI agent chat window and targeted based on 'conversation' context, to enable Amp Free -- $0 version of an otherwise premium product.
Once again, Sourcegraph leading the way on big pricing bets in the age of AI.
Is nothing sacred anymore?
October 15, 2025 at 12:45 AM
Is nothing sacred anymore?