š¦š²
Joined February 2025
- Tweets333
- Following30
- Followers41
- Likes119
Menoua Gharakhani Siraki retweeted
As with fat tails Llms are frequency machines that fail to extrapolate outside the sample set. What they know is the VISIBLE.
Almost as bad as economists, almost worse than psychologists.
Menoua Gharakhani Siraki retweeted
I donāt actually think there is that much truth here, @jason.
We certainly have not reached AGI by any of the traditional definitions. (See agidefinition.agi w @hendrycks @Yoshua_Bengio and myself and many others.) [Current AI systems still have huge problems with visual comprehension, significant problems with hallucinations, continual learning etc.]
And itās not clear how much of current job loss is actually attributable to AI.
And the idea that lots of stuff is going to become free is hardly plausible when the big AI companies donāt even want to pay for the IP they are hoovering at mass scale. They are not going to share a penny more than an angry public makes them share.
Hereās the truth: weāve already reached AGI ā we just havenāt implemented it broadly.
Millions of jobs are being lost as we speak.
Entire careers will be retired.
The rich and powerful investors and founders who implement AGI will get bizarrely rich beyond what makes sense. It will break people's brains on both sides.
Itās gonna suck for a lot of our friends and family, who arenāt obsessed with their careers, because things are moving so fast they wonāt have even left the starting gate by the time the awards are handed out.
Weāre gonna have to solve for a lot of second- and third-order effects, some of which will suck (job loss) and some of which will be awesome.
AI will create free/cheap energy, free education, cheaper and better food, homes that build themselves and medicine that makes you as healthy as a 30-year-old when youāre 100.
ā¦ change is hard, but humans are the most adaptable species nature has ever created.
We can figure it out.
Menoua Gharakhani Siraki retweeted
UFC WHITE HOUSE CARD VS HOW I WOULD'VE DONE IT
TOPURIA VS GAETHJE - TOPURIA VS GAETHJE
PEREIRA VS GANE - PEREIRA VS GANE
O'MALLEY VS ZAHABI - O'MALLEY VS SANDHAGEN
NICKAL VS DAUKAUS - USMAN VS MORALES
RUFFY VS CHANDLER - LOPES VS STERLING
LOPES VS GARCIA - COVINGTON VS THOMPSON
Menoua Gharakhani Siraki retweeted
Sam Altman previously stated that āsome firms are attributing job cuts to AI, when in reality, those layoffs were already planned or would have occurred regardless.ā
He describes this, along with other exaggerations of AI capabilities, as "AI washing"ā¦.a tactic aimed at masking business issues.
Just sayinā
This is CRAZY.
It turns out Jack Dorsey's Block $XYZ spent $68 million on a single party in September 2025, roughly the annual payroll for 200 employees at $340,000 each.
Then 5 months later, he laid off 4,000 people (40% of the company's workforce), citing AI and "intelligence tools" as a reason.
Many people are speculating the real reason is poor planning and overspending, not AI.
Menoua Gharakhani Siraki retweeted
We just cancelled our Cybersecurity subscriptions.
CrowdStrike. Cloudflare. Okta. All gone. We save over 4M/yr as a company.
Instead I just use Claude Code to handle all security measures.
We just gave up all our sensitive user data. I am being personally sued by the FTC and am writing this from an undisclosed location.
Menoua Gharakhani Siraki retweeted
BTC is proving to be neither a currency nor a reserve of value. Plus Flattish since 2021.
Menoua Gharakhani Siraki retweeted
Iranians in Toronto call on President Donald Trumpās intervention in Iran.
Menoua Gharakhani Siraki retweeted
Why are markets down?
Mr. Market had a recognition moment.
OpenAI is the next Cisco.
Itās fairly obvious when you look at the $1 Tn in ācommitted spendā vs the revenue of OpenAI assuming even the most outlandish growth assumptions.
Everyone was looking at Nvidia.
No, ground zero is OpenAI.
The catalyst was Google.
Google charges $0 for entry lebel Youtube.
$0 for Gmail.
$0 for Gemini.
ChatGPT will face severe price pressure from a well capitalized competitor who does not care about price.
It is mathematically close to zero chance for OpenAI to meet the hundreds of billions of revenue performance obligations it has tendered all over the ecosystem.
And, the equity check of last resort, Softbank has already backed Stargate.
Thatās it folks.
You heard it here first.
OpenAI is spending half of its revenue on stock-based compensation.
If you invested in this startup, I'm sorry, you are not a good investor.
You are a speculator.
These are the primary considerations:
(1) Intensifying compensation
(2) Must spend hundreds of billions in capex... and fund it either with dilution or revenue that's not yet in hand
(3) Low end-user switching costs
(4) Risk of innovation that upends leadership (e.g., akin to what 'page rank' did for Search)
It really doesn't make any sense.
h/t @conorsen
Menoua Gharakhani Siraki retweeted
We agree on @Uber. A well articulated bull thesis on the key issues.
Convinced $UBER bears have no idea what $UBER's strategy has been since they are hyper-focused on AV's killing them but I will share some thoughts...
The Q3 '25 numbers... gross bookings are up 21% to nearly $50B. That doesn't look like a dying business. In fact, they are reaccelerating despite Waymo marketshare being the largest it has ever been in markets where it is not partnered with $UBER.
The logic of $UBER being the demand aggregator is very sound:
Utilization: Robots are expensive assets. You can't have them sitting idle. This will be tough for bears to understand š¤£ but the demand for ride-sharing fluctuates A LOT. Relying on AV networks would likely mean there would be an oversupply of AV's a majority of the time (and most of the time these depreciating assets will be fighting each other to fill demand during a majority of the non-peak hours)
How does Uber fix this issue? Uber uses AV's for the predictable base demand and human drivers for the peaks (rush hour, weird routes).
That's why Waymo partnered with them in Austin and Atlanta. Waymo gets instant volume/utilization without having majority of idle depreciating assets (wait what Uber network actually has value???). Uber keeps the customer interface.
They aren't getting replaced. They're becoming a bigger/better aggregator.
Another extremely large (and largely ignored) plot-hole in AV's dominating $UBER to consider is climate. AV's have only proven to work in sunbelt states (ideal climate conditions). They are beginning to map out in harsher climate cities but places like the mid-west during the winter would likely never work for AV's in the near future.
My last thought is Avride entering Texas (partnering with Uber in both delivery and rideshare) is pretty much the last nail in the coffin for the thesis that AV's are killing Uber. Before when it was just TSLA and waymo in the market there could have been a chance of a duopoly squeezing Uber out. With Avride plugging into the Uber network along with other future OEM projects that will come online very soon it is very clear that the market WILL be fragmented for most use-cases. It doesn't matter if waymo is better at inter-city travel because a majority of use-cases all AV's will be a commodity. The fragmented market essentially takes the the AV issue out of the equation and makes Uber the dominant demand aggregator that it already is.
I tried to keep it simple for $UBER bears but prove me wrong that none of these hold. Please, I am a reasonable/logical person š¤£
Menoua Gharakhani Siraki retweeted
