@deserat.bsky.social
Suprised people still haven't learned. If you crawl in bed with Larry Ellison you will get fucked.
December 23, 2025 at 12:54 AM
Suprised people still haven't learned. If you crawl in bed with Larry Ellison you will get fucked.
Genuinely amazed at how bad Pop OS 24.04 is. I've enjoyed Pop OS since it first came out. But 2 corrupted filesystems, and a crazy number of kernel panics later I'm out. Moved back to Ubuntu, repaired the drives and will revisit #PopOS again in a year or two. FYI I'm running a System76 Thelio.
December 22, 2025 at 5:21 PM
Genuinely amazed at how bad Pop OS 24.04 is. I've enjoyed Pop OS since it first came out. But 2 corrupted filesystems, and a crazy number of kernel panics later I'm out. Moved back to Ubuntu, repaired the drives and will revisit #PopOS again in a year or two. FYI I'm running a System76 Thelio.
"AI is America's next Manifest Destiny, and we're ensuring that we dominate this new frontier" -- Emil Michael ( Pentagon Under Secretary )
Hmmm like we dominated ( genocided ) the last one?
Hmmm like we dominated ( genocided ) the last one?
December 11, 2025 at 4:45 PM
"AI is America's next Manifest Destiny, and we're ensuring that we dominate this new frontier" -- Emil Michael ( Pentagon Under Secretary )
Hmmm like we dominated ( genocided ) the last one?
Hmmm like we dominated ( genocided ) the last one?
For coding I find myself throwing out MCP servers and replacing with the CLI tools. That works alot better. Especially the #atlassian mcp which is total garbage.
December 9, 2025 at 10:54 PM
For coding I find myself throwing out MCP servers and replacing with the CLI tools. That works alot better. Especially the #atlassian mcp which is total garbage.
Recent conversation
"That image is perfect." PM
"Thanks we need to pay for the license. It's $200." Designer
Designer goes on vacation.
"Hey what hapend to that image?" Designer
"Client didn't want to pay, we had AI create one very close to it." PM
This is what the future looks like folks.
"That image is perfect." PM
"Thanks we need to pay for the license. It's $200." Designer
Designer goes on vacation.
"Hey what hapend to that image?" Designer
"Client didn't want to pay, we had AI create one very close to it." PM
This is what the future looks like folks.
December 8, 2025 at 5:41 PM
Recent conversation
"That image is perfect." PM
"Thanks we need to pay for the license. It's $200." Designer
Designer goes on vacation.
"Hey what hapend to that image?" Designer
"Client didn't want to pay, we had AI create one very close to it." PM
This is what the future looks like folks.
"That image is perfect." PM
"Thanks we need to pay for the license. It's $200." Designer
Designer goes on vacation.
"Hey what hapend to that image?" Designer
"Client didn't want to pay, we had AI create one very close to it." PM
This is what the future looks like folks.
Reposted
Were ASI to come to fruition, it would be so powerful that it would outcompete human beings in everything – from scientific discovery to strategic warfare. What might happen to our species if we reach this point of singularity, and how can we steer away from the worst outcomes?
In this episode, Nate is joined by AI safety researcher Nate Soares to discuss the dangerous unpredictability of continued ASI development, the “alignment problem,” and the newest safety studies uncovering increasingly deceptive AI behavior.
www.thegreatsimplification.com/episode/203-...
www.thegreatsimplification.com/episode/203-...
December 3, 2025 at 4:02 PM
Were ASI to come to fruition, it would be so powerful that it would outcompete human beings in everything – from scientific discovery to strategic warfare. What might happen to our species if we reach this point of singularity, and how can we steer away from the worst outcomes?
OpenAI should have declared a code red before it lost all of it's top researchers 2 years ago. How it ever thought it could survive that and compete with the company that has been at the vanguard of ML and AI 2 decades is beyond me. Guess they thought they could survive powered only by Altmans ego.
December 4, 2025 at 5:57 PM
OpenAI should have declared a code red before it lost all of it's top researchers 2 years ago. How it ever thought it could survive that and compete with the company that has been at the vanguard of ML and AI 2 decades is beyond me. Guess they thought they could survive powered only by Altmans ego.
Talking with marketing folk and they are already full steam ahead working to enshitify AI. Could ya'll at least wait until it's widely useful?
November 20, 2025 at 7:00 PM
Talking with marketing folk and they are already full steam ahead working to enshitify AI. Could ya'll at least wait until it's widely useful?
Betting all the AI eggs on a single hardware providers basket seems like a very very foolish thing to do. Single Point of Failure anybody?
November 20, 2025 at 5:04 PM
Betting all the AI eggs on a single hardware providers basket seems like a very very foolish thing to do. Single Point of Failure anybody?
I think we've spent to long focusing on making technology easy to use, and not nearly enough time focusing on making the impact of technology easy to understand.
October 28, 2025 at 3:30 PM
I think we've spent to long focusing on making technology easy to use, and not nearly enough time focusing on making the impact of technology easy to understand.
I am endlessly amazed by businesses ability to try any trendy tactic to squeeze every penny out of their customers while completely ignoring the obvious one. Treat your customers well. Treat your loyal customers with loyalty. Did we forget The Clue Train Manifesto?
www.cluetrain.com/book/95-thes...
www.cluetrain.com/book/95-thes...
the cluetrain manifesto - 95 theses
The Cluetrain Manifesto.
www.cluetrain.com
October 27, 2025 at 9:37 PM
I am endlessly amazed by businesses ability to try any trendy tactic to squeeze every penny out of their customers while completely ignoring the obvious one. Treat your customers well. Treat your loyal customers with loyalty. Did we forget The Clue Train Manifesto?
www.cluetrain.com/book/95-thes...
www.cluetrain.com/book/95-thes...
One reason I've always enjoyed programming is, it allows me to see how the codification of ideas affect outcomes, especially at scale. It is one thing to think something is right, it is a very different thing to so what happens when that "right" is applied a million times over. Unexpected Outcomes.
October 7, 2025 at 12:01 AM
One reason I've always enjoyed programming is, it allows me to see how the codification of ideas affect outcomes, especially at scale. It is one thing to think something is right, it is a very different thing to so what happens when that "right" is applied a million times over. Unexpected Outcomes.
If you want to know what a person believes, look at what they do when they are at personal risk. That is the difference between virtue signalling and conviction.
September 22, 2025 at 9:00 PM
If you want to know what a person believes, look at what they do when they are at personal risk. That is the difference between virtue signalling and conviction.
Apparently being anti fascist makes me a terrorist. So be it. So long as nobody accuses me of being ICE or ever committing or supporting an act of violence, I'm good.
September 18, 2025 at 3:23 PM
Apparently being anti fascist makes me a terrorist. So be it. So long as nobody accuses me of being ICE or ever committing or supporting an act of violence, I'm good.
Watching the company I work for create an AI team filled with Business Analysts who have no software background. That team is already "Vibe Coding" solutions for the company. If I wasn't the person who'll have to fix them, the results would be hysterical. Prolly time to find another job.
September 14, 2025 at 12:14 AM
Watching the company I work for create an AI team filled with Business Analysts who have no software background. That team is already "Vibe Coding" solutions for the company. If I wasn't the person who'll have to fix them, the results would be hysterical. Prolly time to find another job.
To think the Democratic Party is going to fight this Fascism is foolish. They could not have worked harder to help create our current situation.
September 10, 2025 at 2:19 PM
To think the Democratic Party is going to fight this Fascism is foolish. They could not have worked harder to help create our current situation.
I find that working with an AI Agent is alot like working with a human. No matter how detailed your ask is the result is a mixture of; some of what I asked for, some things I didn't ask for, some idiocy and some things I learn from. The real difference is that the cycle times are much shorter.
September 3, 2025 at 11:12 PM
I find that working with an AI Agent is alot like working with a human. No matter how detailed your ask is the result is a mixture of; some of what I asked for, some things I didn't ask for, some idiocy and some things I learn from. The real difference is that the cycle times are much shorter.
Given all the benchmarks we have for AI ya'd think there would be a benchmark for evil, or maybe ethics. I'm just spit balling but such a benchmark could include like; emphasis on alignment, copyright violation, use of user data, energy efficiency.
Anybody know of such a thing?
#ai #alignment
Anybody know of such a thing?
#ai #alignment
September 1, 2025 at 9:23 PM
Given all the benchmarks we have for AI ya'd think there would be a benchmark for evil, or maybe ethics. I'm just spit balling but such a benchmark could include like; emphasis on alignment, copyright violation, use of user data, energy efficiency.
Anybody know of such a thing?
#ai #alignment
Anybody know of such a thing?
#ai #alignment
The more I work in AI Coding land the funnier Vibe Coding gets. If the app never leaves your computer, have fun. But oh my, all the problems; security, race conditions, thread safety. There are time and quality improvements to be had - if you are already and expert. But nightmares if you aren't.
August 12, 2025 at 4:22 PM
The more I work in AI Coding land the funnier Vibe Coding gets. If the app never leaves your computer, have fun. But oh my, all the problems; security, race conditions, thread safety. There are time and quality improvements to be had - if you are already and expert. But nightmares if you aren't.
Started Unfollowing accounts simply based on them posting ALOT. It's sad I started following when they were a few well thought out, interesting, or witty posts. But somehow re-posting mass volumes of noise seems to be the new norm. I'm sorry but there weren't 100 things today you needed to share.
July 29, 2025 at 8:16 PM
Started Unfollowing accounts simply based on them posting ALOT. It's sad I started following when they were a few well thought out, interesting, or witty posts. But somehow re-posting mass volumes of noise seems to be the new norm. I'm sorry but there weren't 100 things today you needed to share.
I'm genuinely amazed at how much time Programming agents spend dealing with the linting errors they create. Seems to take longer than the writing the code. Most could simply be fixed by `ruff format`... but then when is charging for token usage...
July 1, 2025 at 1:26 AM
I'm genuinely amazed at how much time Programming agents spend dealing with the linting errors they create. Seems to take longer than the writing the code. Most could simply be fixed by `ruff format`... but then when is charging for token usage...
Show me what Democracy looks like. This is what democracy looks like.
June 14, 2025 at 5:52 PM
Show me what Democracy looks like. This is what democracy looks like.
Seeing a pretty common pattern in LLM's when it comes to doing productivity tasks ( writing code, text analysis ). The more larger the input token set, the smaller the tasks which can be completed satisfactorily. Even latest models get stuck in loops, loose track of where they are, and get sloppy.
June 13, 2025 at 9:51 PM
Seeing a pretty common pattern in LLM's when it comes to doing productivity tasks ( writing code, text analysis ). The more larger the input token set, the smaller the tasks which can be completed satisfactorily. Even latest models get stuck in loops, loose track of where they are, and get sloppy.
Think I might finally be able to drop VSCode for Zed. My computer battery will be very happy about that.
May 7, 2025 at 5:41 PM
Think I might finally be able to drop VSCode for Zed. My computer battery will be very happy about that.