Olivia
Online
Stewart Guenther
@Fit4Fun
Hiker, rain seeker, traveler. Kann Deutsch.
Bay Area
Joined April 2009
355 Following
233 Followers
1.7K Posts
Pou chèf gang yo ak manm gang yo k ap kenbe kat konpatriyòt nou yo an otaj ann Ayiti:
Mesye nou kidnape yo pa konbatan. Yo se teknisyen nan domèn avyasyon ki fè pati misyon nou ann Ayiti, nan kad yon misyon entènasyonal. Misyon nou an gen sèlman pilòt ak pèsonèl teknik, ki gen kòm travay prensipal fè evakyasyon medikal epi sove lavi moun, pa fè lagè.
Men, nou te kidnape yo yon fason lach pandan yo t ap retounen apre yo te fin achte manje nan yon makèt nan zòn nan.
Se poutèt sa, jodi a nou te lanse yon operasyon militè pou pote yo sekou. Nou gen yon sèl objektif: fè kat konpatriyòt nou yo retounen lakay yo vivan, san okenn mal.
Men, mwen vle pou nou menm, chèf gang yo, konprann yon bagay byen klè: atak jodi a pa anyen devan sa k ap vini si nou pa remèt toule kat otaj yo vivan.
Epi pa pran pawòl nou yo pou menas san fondman. Nou deja fini ak gang yo nan El Salvador, e nou kapab fini ak nou tou.
Nou toujou gen tan pou evite sa.
Remèt nou kat konpatriyòt nou yo vivan, san okenn mal, epi tout bagay ap retounen jan sa te ye anvan kidnaping lan.
Desizyon an nan men nou.
🙏🏼🥴😞
Bryson fire in Malibu is looking bad.
Prayers up for all the residents.
💯
Please take a moment to watch this before asking, yet again, “Will this all go away if Trump loses the midterms and Republicans lose in 2028?”
Most games fail. I made ~$2 million in my solo indie gave dev career which puts me in the top 0.1% of solo indie game devs. You know how many of my games I consider a financial success?
One. 80% of that revenue was one game. Paid for the rest. If you're in software for the money, game dev is not the way to go. Divide my revenue up over the hours I worked, subtract expenses, and I made about what the average junior software engineer makes working for a big company. A third of what I make now.
Make games because you love to make games. Make them with a.i. if you love a.i. Make them without a.i. if you enjoy that. But know that you've picked a market where your odds of financial success are about as low as they possibly can be for a software engineer whether you use a.i. or not.
So many anti-ai haters are throwing stones in a glass house. The amount of "indie game dev" labels I see on people's profiles who have never shipped a game in their life is absurd. Ship a game first, then talk shit. But know that you're talking shit, not speaking the truth. I don't know what motivates people who have never shipped a game in their life to tell other people that their game will fail, but if you're on the other end of this, just know that they have no idea what they are talking about. Nobody does.
Gamedev is even less understood than filmmaking and 80% of the 600 films that hit theatres every year are financial flops. 99% of advice in gamedev is complete bullshit. Millions of games get made, some of them succeed, then the successful devs give talks at GDC like they knew what they were doing. Then a bunch of people copy that formula and fail. There is no formula. Do your own thing. Explore. Discover something nobody else has discovered. Create a new mechanic. Go crazy. And enjoy the process.
👀 signal 👍
I spent 48 hours in SF hanging out with billionaires, data labs and AI founders and execs running $1B to $100B companies.
My observations:
1. Sooo MANY stories of data brokers making like $2M a WEEK selling company data to labs. I can't tell if it's the beginning or a bubble, and I kept wondering which niches are left. Digging into it actively.
2. At dinner, someone asked the table, "If a new model drops next month, does your company get better or worse?" HALF the table went quiet! I think that's the best question to ask about any AI startup right now.
3. The bar people talk about now is hitting $1B in revenue in 12 months (I guess that's the new unicorn?). What I didn't hear much about is retention. AI revenue can come in fast and leave fast, and I think we're going to see a lot of stories of AI companies with huge revenue and terrible retention.
4. It's pretty clear we're living through the reinvention of the computer. The model is the CPU, the harness is the operating system, and skills and plugins are the apps. Thinking about it that way makes it easier to see where the whitespace is.
5. We're maybe 12 months from an agent that can watch you work for a week and then do your job without any instructions. The screen recording plus agent memory plus local model combination makes this possible right now.
6. Agent harnesses feel like the new GPT wrappers, in a good way, especially vertical ones. The best teams I talked to treat the model as swappable. One founder walked me through routing different models to different jobs: one for taste and visuals, a cheaper one for the everyday stuff, another for text.
They use prompts they can move across models over fine-tuning, and A/B test the output on what users actually do with it: do they edit it, export it, share it? A vertical harness is basically four things: the connectors for that industry's data, evals written by people who know the work, rules for what the agent can and can't touch, and a clean screen for a human to review. The model is maybe 20% of it.
7. There's a lot of reverse engineering going on/cloning going on more than I've ever seen. People are decompiling software and rebuilding it with agents, and someone showed a suite of like 10 apps that are worth $5B+ each all that he cloned open source. Me and everyone are still thinking through what it means when software gets this easy to copy.
8. My guess is the value moves to data. If the software can be rebuilt quickly, what's worth paying for in a legacy SaaS company is the years of customer data and workflows inside it. Some SaaS companies are secretly data companies.
Gets me thinking about how do you truly value data for this saas companies? Like if we were buying these dying saas companies at 2x ARR before (as example), maybe there is a better metric. Idk, just thinking out loud here.
9. It feels like a good chunk of legacy software won't make it to the agentic era, especially anything built around people clicking through a lot of screens.
10. Every company I talked to wants to pipe its internal data into one place and expose it to agents through MCP. Then they look at the data and realize it's a mess: five versions of the same customer, old Slack channels, PDFs nobody can search. Cleaning that up feels like a big consulting opportunity!
11. A lot of founders are building the basic building blocks for agents: APIs, data layers and tools. Some of the companies I saw barely have a UI because the main user is an agent. I really believe the startups that build primitives to this agentic era will do extremely well. Building blocks is the best way to think about it.
12. Labs can get plenty of training data, but it's harder to get real examples of agents failing inside companies. Whoever has those logs has something valuable.
13. Someone running a big consumer company told me the next fight is over memory. The model that remembers the most about you wins, and the smartest model matters less than people think.
14. Physical AI came up a lot. Astra I think was the first model that got people into AI can actually do CAD, and I think a lot of people are thinking what does vibe manufacturing really look like. The line I keep thinking about came from a robotics founder: the hardest problems in AI are deployment problems now, not research problems. Their robots work on their own, call a human when they get stuck, and every one of those calls becomes training data.
15. Cybersecurity companies are having a moment and are actually cool companies to build now, especially the ones that clean up after a breach. Every agent with access to your systems is a new door into your company, and a lot of those doors were left open.
16. Some startups are worth more dead than alive now. A shutdown company with 5 years of Slack, email and code can get paid for licensing all of it. I wouldn't be surprised if founders start factoring that into when they shut down.
17. People are starting to talk about monetizing the agent web. It feels really early, but when the most popular user of the internet is an agent, what startups should you be building?
18. I think agent insurance is coming. Once an agent makes an expensive mistake, insurers will want audits before they cover anyone.
19. Several people said the hire they want most is someone with taste who can write evals. The models are getting good at generating ideas, so the hard part is judging which ones are good.
20. New trend I observed: startup raises VC, uses venture dollars to become a creator, startup fails, founder falls back on being a full-time creator and incubates other products bootstrapped.
21. Getting SIGNIFICANT distribution is HARD right now. Founders are sponsoring buses, buying billboards like crazy and throwing parties with famous DJs to get noticed. Kinda feels dotcom era in that sense.
Attention is what's scarce! The growth guy going viral sharing his UGC strategy from Wispr Flow is a good example. It's the sauce everyone wants, and I truly believe there are probably only 50 people on the planet who know how to scale these types of businesses efficiently in this era.
Always good to get that dose of SF. I felt sooo lucky to have such an incredible trip during such an incredible time in building businesses. Surreal in so many ways.
Happy to be home and locked in.
Until next time!
👀
The guy who invented Amazon Prime disappeared for six years. He came back this week with $250 million and a factory that turns lines of code into tiny working machines. Type different code, a different machine comes out the other end.
That sounds fake. Here's why it took six years.
A chip fab is the most precise factory humans have ever built, and it makes exactly one kind of thing. Flat silicon that processes information. The moment you want moving parts, multiple materials, or real 3D structure, you leave the fab for a world where getting one new device into production takes years and tens of millions of dollars. Engineers call it one product, one process. It's the reason micro-machines never got their Moore's Law.
Holden's Matter Compiler kills the per-product process entirely. Features down to a few microns, no new tooling, and the system measures and corrects every step while it builds, then feeds what it learned into the next design.
His first product is the tell on how sharp the strategy is.
PrimeSwitch is a 150-amp relay for 800-volt power in AI data centers. It opens in roughly 50 microseconds, about 1,000x faster than a normal contactor. So the first thing to come out of a universal factory is aimed at the one market where demand is basically infinite and a faster switch means less fried hardware.
Neal Stephenson wrote the Matter Compiler as fiction in 1995. Holden just shipped the first real product out of one, and it's a light switch for AI data centers.
T-Mobile spent five years trying to get rid of the exact spectrum Elon just called "the last critical piece" of complete phone coverage in America.
This 800 MHz block is Sprint's old Nextel spectrum, the stuff that powered push-to-talk walkie-talkie phones in 2004. The DOJ forced T-Mobile to divest it as a condition of the $26 billion Sprint merger. Dish had first rights at $3.59 billion and walked away because it couldn't raise the money. T-Mobile finally offloaded it to an investment firm called Grain Management last year for $2.9 billion plus a spectrum swap.
Grain's original plan was selling it to utilities for private networks.
Fourteen months later, SpaceX buys the whole portfolio, and shares of the carriers who passed on it fall as much as 8% in a single afternoon.
Here's why it matters so much. Starlink's satellite signal at 2 GHz can reach your phone outdoors, but it dies at the first wall. 800 MHz penetrates buildings. Most phones already have the band built in. So satellites handle the open sky, this low-band layer handles indoors, and suddenly Starlink Mobile covers everywhere a normal carrier does without the tower network.
Stack the moves. $17 billion for EchoStar's 50 MHz of 2 GHz spectrum. FCC approval for 15,000 Gen2 satellites this week. Now the low-band coverage layer.
Verizon, AT&T, and T-Mobile paid $81 billion in a single auction for C-band spectrum in 2021, then spent years and tens of billions more deploying it on towers. SpaceX assembled a nationwide carrier out of a distressed-asset sale, a forced divestiture castoff, and rockets it already owns.
The walkie-talkie spectrum from 2004 ended up being the piece that lets satellites replace cell towers. Nobody who touched it for twenty years saw that coming.
😅
While Hollywood is busy figuring out how to reinvent Bond for modern audiences, I decided to do something different: Make Bond Great Again.
Meet Agent 047. Because sometimes the most obvious casting choice is also the most tremendous one.
Episode 2 of TEHRAN ROYALE is coming! And this teaser trailer builds up the hype even more! A James Bond parody series starring Donald Trump as Agent 047. Vintage spy glamour, dangerous company, lots of car and bike chases, and a man who has never doubted himself for a second.
Starring: Melania Trump as Agent Melatonin
Elon Musk as Q
Donald Trump as Agent 047
And for the first time, Zohran Mamdani as a known jihadist terrorist
Created with @higgsfield_ai- an amazing platform, that gives you all the necessary instruments for all stages of AI video creation.
I understand why people want research to be openly accessible, I do too. But independent research still has real costs. I’ve already spent thousands on cloud compute building the EM solver, and the next system I’m looking at is around $15k. I’m doing this around a full time job, conference presentations, and homeschooling my kids. Universities have grants, institutional compute, and paid research time. If we want serious independent research to exist, it needs a way to fund itself too.
Labor Secretary Keith Sonderling announces 72,000 manufacturing jobs added this year. Including 9,000 in September alone.
September Jobs Report:
Construction: +11,000
Healthcare: +17,000
Private sector: +46,000
Government: -17,000
Manufacturing has expanded for nine straight months after years of decline. Private-sector jobs are growing while government employment is shrinking.
This is what I voted for.
👀
a 67-year-old retired us navy captain trained an llm in his winnebago that beat google deepmind, nvidia, and noaa at weather forecasting, and sam altman, jensen huang, elon musk, and demis hassabis are all now trying to buy it
his name is carl reynolds. he lives in his rv full-time between kohler, wisconsin and st. george, utah.
31 years as a naval meteorologist. he called the 2018 typhoon jelawat track 18 hours before joint typhoon warning center did.
retired in 2022, bought a 2019 winnebago, started training the model on his laptop
by 2025 it ran on a $2,400 rack of gaming pcs bolted to the inside of the rv, cooled by a modified home ac unit, drawing 900 watts from his lithium bank
he calls it stratus-7
stratus-7 beat google deepmind's graphcast-2 on 96-hour precipitation across the continental us in august, september, and october 2025. beat nvidia earth-2 on hurricane track for atlantic season 2025. beat both on jet stream position by day five.
he posts free forecasts to a public github repo every six hours
demis hassabis linked to it in an internal deepmind slack in january
sam altman's chief of staff emailed carl in february offering "a consulting arrangement"
jensen huang shipped him 8 h200s to the walmart parking lot in march with no note and no invoice
carl sold the h200s on ebay. donated the $340k to a naval widows fund. posted the receipt.
that is when the dod noticed
palantir approached in april and offered $12m
lockheed offered $18m plus a defense clearance
raytheon offered $22m and equity in a spinoff
elon musk offered $35m plus a permanent xai research seat over dm on september 4
carl said no to all of them
on september 17 he received an unsigned letter delivered to his rv while parked at a walmart in cheyenne, wyoming. the letter offered $80 million from an unnamed federal buyer and included one phone number.
he did not call.
on september 20 his rv was pulled over by wyoming state troopers for a broken taillight. it was not broken. they searched for four hours. they found nothing they were allowed to keep.
on september 23 his github repo was locked "pending review" by github trust and safety. he mirrored the last six forecasts to a fediverse instance in iceland.
on september 26 the noaa director resigned "for personal reasons"
on september 28 sundar pichai tweeted a hurricane forecast image with no caption. every meteorologist on the platform recognized the color palette. it was stratus-7's.
on september 29 carl posted a screenshot of a gmail draft addressed to demis hassabis. subject line: thanks, that was cool, still not for sale.
carl is currently parked at a truck stop outside boise. the rv has 187,000 miles. 3.4 million followers on the fediverse. no lawyer.
when a journalist reached him yesterday and asked why he will not sell, he said one thing
i was navy for 31 years. weather saved my life four times. it does not belong to a company.
yann lecun quote-tweeted it at 2:14am with three words
the correct answer
👀👀👀
This is for all the marbles, people... https://t.co/v723vdRhYv
The most “radical” thing a man can do in today’s culture is find a godly woman, marry her, be faithful to her, and raise children together. Creating broken homes doesn’t suddenly become empowering just because you dress it up in faux revolutionary terms.
👀
I had the opportunity to go through the physics behind Havenbound’s simulation in detail.. the axis calculations, energy, timing, and surface speeds.
After working through the numbers with him, I fully endorse the physics in the app and can’t wait for launch on Thursday!
If you have <$1M, the only investing advice you need is:
Learn how to make money not invest it.
Both are hard. But it’s easier to turn $100k into $2M actively than it is to turn $100k into $500k passively.
Besides, you don’t wanna save for 20 years only to realize inflation compounds too.
The Pittsburgh Steelers asked an Englishman to sing the American national anthem.
Bruce Dickinson did it a cappella, no autocue, having never sung it before in his life.
He held "free" for six seconds.
Last Seen Users on Sotwe
Most Popular Users
Elon Musk
@elonmusk
241.5M followers
Barack Obama
@barackobama
118.8M followers
Cristiano Ronaldo
@cristiano
114.5M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.1M followers
Rihanna
@rihanna
98.7M followers
NASA
@nasa
92.3M followers
Justin Bieber
@justinbieber
91.8M followers
KATY PERRY
@katyperry
90M followers
Taylor Swift
@taylorswift13
84M followers
Lady Gaga
@ladygaga
75.5M followers
Virat Kohli
@imvkohli
73.5M followers
Kim Kardashian
@kimkardashian
70.9M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
66.5M followers
Bill Gates
@billgates
65.3M followers
Selena Gomez
@selenagomez
63.1M followers
The Ellen Show
@theellenshow
62.2M followers
CNN
@cnn
61.7M followers
X
@x
60.7M followers