Troy Wolverton
@troywolv.bsky.social
Longtime tech and Bay Area journalist. Currently a senior tech correspondent with The San Francisco Examiner. https://www.sfexaminer.com/users/profile/troy_wolverton/
Reposted by Troy Wolverton
If I were editing a column about Charlie Kirk’s political genius, I would tell the writer he needed to quote Kirk’s actual political beliefs. If the writer responded that doing so would weaken or distract from his argument, it’d be a pretty good sign that the column was not ready for prime time.
September 11, 2025 at 2:47 PM
If I were editing a column about Charlie Kirk’s political genius, I would tell the writer he needed to quote Kirk’s actual political beliefs. If the writer responded that doing so would weaken or distract from his argument, it’d be a pretty good sign that the column was not ready for prime time.
Venture capitalists responded to George Floyd's 2020 death by stepping up their investments in Black-founded startups, according to a new study. But in retrospect, what happened ended up looking a lot like "tokenism," the researchers say.
Study: VCs engaged in ‘tokenism’ after Floyd death
Black-led startups saw surge in funding, then drop.
www.sfexaminer.com
September 2, 2025 at 11:15 PM
Venture capitalists responded to George Floyd's 2020 death by stepping up their investments in Black-founded startups, according to a new study. But in retrospect, what happened ended up looking a lot like "tokenism," the researchers say.
In a recent study of the reasoning abilities of generative AI systems, a team of researchers at Arizona State concluded that large language models like OpenAI's GPT-5 don't really reason at all. Their study builds on other recent research that casts doubt on AI's reasoning abilities.
Researchers: AI reasoning a ‘brittle mirage’
Models fail to solve problems outside their training, study contends.
www.sfexaminer.com
August 26, 2025 at 5:17 PM
In a recent study of the reasoning abilities of generative AI systems, a team of researchers at Arizona State concluded that large language models like OpenAI's GPT-5 don't really reason at all. Their study builds on other recent research that casts doubt on AI's reasoning abilities.
Reposted by Troy Wolverton
media doesn't have to pretend that every statement from Donald Trump, a serial liar with a long record of proven falsehoods, has to be evaluated in a vacuum.
August 26, 2025 at 2:29 AM
media doesn't have to pretend that every statement from Donald Trump, a serial liar with a long record of proven falsehoods, has to be evaluated in a vacuum.
The yearslong battle over public access to the Bay Area's Martins Beach — via a road venture capitalist Vinod Khosla tried to close — will continue on for at least a little while longer.
Martins Beach case delayed again for negotiations
Lawyers for state, VC Khosla sought more time.
www.sfexaminer.com
August 21, 2025 at 10:18 PM
The yearslong battle over public access to the Bay Area's Martins Beach — via a road venture capitalist Vinod Khosla tried to close — will continue on for at least a little while longer.
This is a very good piece by Miriam Jordan of @nytimes.com. But it's frankly journalistic malpractice at this point to not call Trump's lies just that.
Instead, we get Jordan saying Trump "amplified a baseless claim by his running mate, JD Vance, that Haitians there were stealing and eating pets."
Instead, we get Jordan saying Trump "amplified a baseless claim by his running mate, JD Vance, that Haitians there were stealing and eating pets."
An Ohio City Faces a Future Without Haitian Workers: ‘It’s Not Going to Be Good’
www.nytimes.com
August 21, 2025 at 7:53 PM
This is a very good piece by Miriam Jordan of @nytimes.com. But it's frankly journalistic malpractice at this point to not call Trump's lies just that.
Instead, we get Jordan saying Trump "amplified a baseless claim by his running mate, JD Vance, that Haitians there were stealing and eating pets."
Instead, we get Jordan saying Trump "amplified a baseless claim by his running mate, JD Vance, that Haitians there were stealing and eating pets."
California Asm. Rebecca Bauer-Kahan (@rbkforassembly.bsky.social) has been trying for three years to push through a bill to offer some protections against discrimination by automated decision systems. I took a look at the fight over her latest effort — AB 1018.
Bill regulating AI decision systems draws fire
Business groups prioritizing defeating California AB 1018.
www.sfexaminer.com
August 20, 2025 at 5:24 PM
California Asm. Rebecca Bauer-Kahan (@rbkforassembly.bsky.social) has been trying for three years to push through a bill to offer some protections against discrimination by automated decision systems. I took a look at the fight over her latest effort — AB 1018.
Reposted by Troy Wolverton
"How bad slavery was" is worse. Much worse. Much worse than you think, much worse than you're taught, much worse than museums depict. Worse.
August 19, 2025 at 7:42 PM
"How bad slavery was" is worse. Much worse. Much worse than you think, much worse than you're taught, much worse than museums depict. Worse.
In talking with family and friends — even people who are very technologically literate — it's pretty clear that many don't understand how systems like ChatGPT work or their limitations and risks. So, I put together a guide — a "Things to Think About" when using generative AI chatbots.
Keep these ChatGPT limitations, risks in mind
Remember these tips on how chatbots work, fall short.
www.sfexaminer.com
August 19, 2025 at 10:29 PM
In talking with family and friends — even people who are very technologically literate — it's pretty clear that many don't understand how systems like ChatGPT work or their limitations and risks. So, I put together a guide — a "Things to Think About" when using generative AI chatbots.
Reposted by Troy Wolverton
As I said to @troywolv.bsky.social, don't use chatgpt when you need to be right about something. It's not, in fact, an all-knowing orb. (Fun convo, fun piece) www.sfexaminer.com/news/technol...
Keep these ChatGPT limitations, risks in mind
Remember these tips on how chatbots work, fall short.
www.sfexaminer.com
August 19, 2025 at 9:12 PM
As I said to @troywolv.bsky.social, don't use chatgpt when you need to be right about something. It's not, in fact, an all-knowing orb. (Fun convo, fun piece) www.sfexaminer.com/news/technol...
The widespread disappointment with OpenAI's new GPT-5, sparked in part by it making similar mistakes as its predecessors, highlighted the gulf between the massive hype that's being used to sell these generative AI systems and their actual capabilities.
GPT-5 debut spotlights AI’s hype-reality gap
OpenAI’s “PhD-level” intelligence made dumb errors.
www.sfexaminer.com
August 14, 2025 at 6:10 PM
The widespread disappointment with OpenAI's new GPT-5, sparked in part by it making similar mistakes as its predecessors, highlighted the gulf between the massive hype that's being used to sell these generative AI systems and their actual capabilities.
While some of the biggest companies in tech have put in place mandatory in-office policies, many large San Francisco tech companies are doing the opposite. I took a look at how local tech companies are now thinking about in-office and remote work.
Many SF tech companies sticking with remote work
Dropbox, Pinterest, Figma and more remain committed.
www.sfexaminer.com
August 12, 2025 at 6:52 PM
While some of the biggest companies in tech have put in place mandatory in-office policies, many large San Francisco tech companies are doing the opposite. I took a look at how local tech companies are now thinking about in-office and remote work.
Reposted by Troy Wolverton
The phrase “ parents have a legitimate concern about integration of races in schools and as a party, we haven’t done a very good job of listening to those concerns.” would shock and offend I would hope. A similar statement about trans kids should also shock and offend.
August 11, 2025 at 12:10 PM
The phrase “ parents have a legitimate concern about integration of races in schools and as a party, we haven’t done a very good job of listening to those concerns.” would shock and offend I would hope. A similar statement about trans kids should also shock and offend.
This is heartrending and horrifying — and oh, so human.
July 11, 2025 at 5:43 PM
This is heartrending and horrifying — and oh, so human.
Reposted by Troy Wolverton
Far from achieving some kind of super intelligence, even the most advanced generative AI systems — the so-called “reasoning” or “thinking” models — struggle to reason beyond a certain level of complexity.
Studies: AI ‘reasoning’ models might not be
Artificial intelligence struggles with more complex, novel problems.
www.sfexaminer.com
July 5, 2025 at 5:59 AM
Far from achieving some kind of super intelligence, even the most advanced generative AI systems — the so-called “reasoning” or “thinking” models — struggle to reason beyond a certain level of complexity.
Reposted by Troy Wolverton
Elon Musk casually, and illegally, dismantled USAID. According to a new study in the Lancet, that action will lead to *14 million* premature deaths between 2025 and 2030.
Again: Elon Musk is directly responsible for 14 million coming deaths. 4.5 million will be children.
Again: Elon Musk is directly responsible for 14 million coming deaths. 4.5 million will be children.
Evaluating the impact of two decades of USAID interventions and projecting the effects of defunding on mortality up to 2030: a retrospective impact evaluation and forecasting analysis
USAID funding has significantly contributed to the reduction in adult and child mortality
across low-income and middle-income countries over the past two decades. Our estimates
show that, unless the a...
www.thelancet.com
July 2, 2025 at 8:00 PM
Elon Musk casually, and illegally, dismantled USAID. According to a new study in the Lancet, that action will lead to *14 million* premature deaths between 2025 and 2030.
Again: Elon Musk is directly responsible for 14 million coming deaths. 4.5 million will be children.
Again: Elon Musk is directly responsible for 14 million coming deaths. 4.5 million will be children.
Reposted by Troy Wolverton
A political movement driven by reasonable concerns about sports and medicine would have no problem admitting that trans people face discrimination. The fact that "gender criticals" constantly deny basic realities is sign 1,087 that they are in fact driven by bigotry.
July 2, 2025 at 2:38 PM
A political movement driven by reasonable concerns about sports and medicine would have no problem admitting that trans people face discrimination. The fact that "gender criticals" constantly deny basic realities is sign 1,087 that they are in fact driven by bigotry.
The rapid development and adoption of AI technologies has resulted in huge demand for electricians. I spoke with union leaders and members in different parts of the country about what they're seeing — and how electricians in ground zero of the AI boom aren't benefiting from it.
AI boom boosting electrician demand — outside SF
In AI’s ground zero, electrical work hard to find.
www.sfexaminer.com
June 10, 2025 at 6:19 PM
The rapid development and adoption of AI technologies has resulted in huge demand for electricians. I spoke with union leaders and members in different parts of the country about what they're seeing — and how electricians in ground zero of the AI boom aren't benefiting from it.
Reposted by Troy Wolverton
The debate over AI safety didn't end with California Gov. Gavin Newsom's veto of last year's controversial Senate Bill 1047 from state Sen. Scott Wiener. A new bill that's essentially a derivative of SB 1047 is progressing through New York's legislature.
NY AI safety bill picks up from Wiener’s
Bill author tries to apply lessons from SB 1047 fight.
www.sfexaminer.com
June 2, 2025 at 5:44 PM
The debate over AI safety didn't end with California Gov. Gavin Newsom's veto of last year's controversial Senate Bill 1047 from state Sen. Scott Wiener. A new bill that's essentially a derivative of SB 1047 is progressing through New York's legislature.
AI boosters argue we shouldn't worry about it boosting carbon emissions, because future AI will be "fixing the climate," as Sam Altman put it. But critics say boosters are blurring the lines between what they're developing and what might really help in the climate fight.
Why ChatGPT is unlikely to solve climate change
Critics say boosters conflating different types of AI.
www.sfexaminer.com
May 28, 2025 at 5:29 PM
AI boosters argue we shouldn't worry about it boosting carbon emissions, because future AI will be "fixing the climate," as Sam Altman put it. But critics say boosters are blurring the lines between what they're developing and what might really help in the climate fight.
Reposted by Troy Wolverton
very cool that for ten years people at the heights of political commentary insisted to the point of rage that the paramount threat to free speech was “censorious” college students
after announcing that foreign students are being banned from Havard, Noem warns "this should be a warning to every other university to get your act together."
May 22, 2025 at 9:44 PM
very cool that for ten years people at the heights of political commentary insisted to the point of rage that the paramount threat to free speech was “censorious” college students
Salesforce used to loudly tout its climate goals, investments and progress towards reducing its emissions. But when it recently made a change to its emissions targets — retreating from two key ones — it was much less vocal about it.
New Salesforce climate goals quietly lower bar
Company no longer promising absolute emissions drop.
www.sfexaminer.com
May 22, 2025 at 6:52 PM
Salesforce used to loudly tout its climate goals, investments and progress towards reducing its emissions. But when it recently made a change to its emissions targets — retreating from two key ones — it was much less vocal about it.