Pat MacNeill

4.4K posts

Pat MacNeill banner
Pat MacNeill

Pat MacNeill

@MacNeillPatrick

انضم Eylül 2012
1K يتبع308 المتابعون
Pat MacNeill أُعيد تغريده
Polymarket
Polymarket@Polymarket·
JUST IN: Trump pledges not to annex Canada & turn it into the 51st state out of respect for King Charles III.
English
241
161
3.1K
223.1K
Pat MacNeill
Pat MacNeill@MacNeillPatrick·
@MagnusSigurdss $UPST is now a cheap stock on fundamental value of its earnings. So the stock will start to attract value investors. Momentum/tech investors will come back only after it is up 100% from here.
English
0
1
4
184
Magnus Sigurdsson
Magnus Sigurdsson@MagnusSigurdss·
Fundamentals of $UPST keep improving, origination per day at record, as well as origination volume, in a challenging macro environment. You would think the stock should be well, yet the stock has gone nowhere in this period, despite wild swings up and down. Market seems to be solely focused on credit risk, and $UPST as a balance sheet lender, giving 0 weight to it's tech. With time it should be proven wrong
Magnus Sigurdsson tweet media
English
1
0
12
751
Pat MacNeill أُعيد تغريده
Diogenes of Sinope
Diogenes of Sinope@DiogenisSinopis·
Préveli Beach, also known as Palm Beach, is a popular beach in the Rethymno region on the southern coast of Crete in Greece. It's famous for its unique palm forest bordering the beach and the turquoise, crystalline waters of the Cretan Sea. 💙🇬🇷🌴🏞️🌴🇬🇷🩵 en.wikipedia.org/wiki/Preveli_B…
English
0
34
253
9.5K
Pat MacNeill أُعيد تغريده
Megalithic Mysteries
Megalithic Mysteries@Megalithic12000·
🚨 Japan placed stone warnings along its coastline centuries ago. In 2011, the ocean proved them right. Scattered along Japan's coast are hundreds of ancient markers known as tsunami stones. Some are over 600 years old. They carry one message: do not build below this point. 🔹Aneyoshi obeyed and survived 2011 🔹Some stones are over 600 years old 🔹Hundreds placed along Japan's coast 🔹Over 20,000 killed in the 2011 tsunami 🔹Towns below the markers were destroyed One stone in Aneyoshi reads "Remember the calamity of the great tsunamis. Do not build any homes below this point." The village obeyed. When the 2011 tsunami hit, the water stopped just below the stone. Aneyoshi survived. Surrounding towns that built below the markers lost over 20,000 people. A Tohoku University professor said "It takes about three generations for people to forget." Survivors carved warnings into stone because they knew memory fades. Civilisation after civilisation did the same thing thousands of years earlier. If the Japanese stones proved deadly accurate after 600 years, why do we assume the older warnings are just myths?
Megalithic Mysteries tweet mediaMegalithic Mysteries tweet media
English
181
4.8K
33.3K
3.2M
Pat MacNeill أُعيد تغريده
𝐓𝐗𝐌𝐂
𝐓𝐗𝐌𝐂@TXMCtrades·
"Open the fuckin Strait" I whisper to my wife as I snuggle up beside her
English
1K
4.8K
66.8K
2.4M
Pat MacNeill
Pat MacNeill@MacNeillPatrick·
@HenryInvests 60% growth in originations which laps a 89% growth Quarter and its trading at a lower PE multiple of a regional bank stock...
GIF
English
1
0
6
391
Pat MacNeill أُعيد تغريده
Henry
Henry@HenryInvests·
Upstart $UPST March originations come in at $1,263M, up 17% MoM, 60% YoY and crushing my $1,154M estimate. Woah. 🎤 🔥
Henry tweet media
English
9
5
127
12.7K
Pat MacNeill
Pat MacNeill@MacNeillPatrick·
@HenryInvests This is an incredible story. Upstart continues to accelerate growth, funding partners continue to demand more of their loans despite problems in other parts of the private credit market. And yet selling pressure in the stock is relentless.
English
1
0
4
259
Henry
Henry@HenryInvests·
@MacNeillPatrick My initial hunch as well. March originations are 8% higher than December (best seasonal month) - that is 2x the growth rate of last year. Auto and HELOC appear to be cooking
English
1
0
11
643
Pat MacNeill أُعيد تغريده
Martin Varsavsky
Martin Varsavsky@martinvars·
In 2005 I built Fon on a simple idea: millions of WiFi routers sit idle most of the day. Share that spare capacity and you build a global network without laying a single cable. Millions of homes joined. Telcos partnered. No infrastructure needed. Now Jack Dorsey’s Block just launched mesh-llm. Same idea, applied to AI. Pool your idle GPU and suddenly a group of people with modest machines can run large open models that none of them could run alone. Models split automatically across nodes. No cloud provider, no API fees, no one controls your data. The timing matters. Google DeepMind released Gemma 4 today under Apache 2.0. A 31B model that competes with much larger closed models. A 26B MoE variant that only activates 3.8B parameters at inference. Edge models that run on a phone. All free to download, free to use commercially, no restrictions. Combine mesh-llm with Gemma 4 and you get the Fon of AI. Distributed compute running frontier open models. No central server. No per-query cost. Total privacy. The intelligence stays with the people who pool the hardware. Twenty years ago the scarce resource was connectivity. Today it is compute. The solution is the same: share what you have, access what you need.
jack@jack

mesh-llm: pool compute to run open models. built by @michaelneale at block: docs.anarchai.org

English
40
89
600
130.9K
Pat MacNeill أُعيد تغريده
Micke Andersson
Micke Andersson@TetsuoIronman·
@policytensor The Arabian plate is converging with the Eurasian so the Hormuz strait will close by itself in a few millions years effectively rendering this entire issue irrelevant.
English
68
1.3K
13.1K
523.7K
Pat MacNeill
Pat MacNeill@MacNeillPatrick·
$UPST Q1 is projected to show a Q/Q increase in Revenue according to @HenryInvests. The key for Q1 will be continued positive cash flows from operations and investing. With the recent funding announcements I expect that to be the cash and for loans on BS to continue to decline.
Henry@HenryInvests

⚡️Preliminary UpstartIQ March Originations and 1Q26 Revenue Estimates #UTPI / $UPST Upstart will be posting March origination data this Friday (the 3rd of the month). I believe March originations will come in ~ $1,154M, up 7.3% MoM, up 45.6% YoY. Here's how I arrive at that estimate (view graphics): 1) Originations ($M) per day is simply monthly origination volume / number of days in the month (this is different from the normalized metric Upstart provides) 2) I estimate that Originations ($M) per day will grow directly in line with the growth of Trustpilot Reviews per day. 3) I then apply this growth rate in MoM reviews per day (-3.1% from Feb to Mar) to Originations ($M) per day and multiply it by the 31 days of March. That puts me at $1,154M in expected origination volume for March. Obviously, the 1:1 ratio/relationship between Trustpilot reviews per day growth and Origination ($M) per day growth is overly simplistic, so please take that with a grain of salt. It did prove rather accurate in February, however ($1,069M est. vs $1,075M actual). Now that I have all of the monthly origination volume for 1Q26 (the reported January and February numbers, plus my March estimate), I can attempt to calculate Upstart's take rate per month by dividing monthly fee revenue by monthly originations. Because Upstart does not provide fee revenue by month (instead only providing it for the quarter), each month's revenue from fees is a pro rata share of the respective quarter's fee revenue. This is not a perfect methodology. monthly fee revenue / monthly origination volume = take rate %. You can see that take rate has been trending down (11.9% in July 2024 -) 7.6% December 2025), reflecting management's commentary towards the intentional dynamic of reducing take rates to achieve larger market share, pass savings along to the borrower, plus the dynamic of lower margins for new products Auto & HELOC. For 1Q26 (January - March) I had previously assumed that Take Rate would trend down slightly from the already subdued level of December 2025. I think that was too overly conservative. While December's take rate was 7.6%, the blended average for 4Q25 was still 8.3%, as compared to 9.1% in 3Q25, 8.6% in 2Q25, and 8.7% in 1Q25. For 1Q26, I am choosing to apply a blended 8.07% take rate across all three months in the quarter. This represents a ~63 bps decline from the 8.7% blended average of 1Q25, or a decline of 7.2% YoY. When you apply my take rate assumptions (8.1% Jan, 8.1% Feb, 8.0% Mar), you arrive at monthly fee revenue taken as take rate * monthly originations ($). When adding those estimates together, I am arriving at a revenue from fees estimate for Upstart in 1Q26 at ~ $269M as compared to $265M in 4Q25. Remember, the first quarter of the year is always seasonal and almost always has lower revenue than that of the fourth quarter. As for net interest income, Upstart has guided for $100M for 2026. My assumption (a pure guess) is that NII will look something like $31M -) $27M -) $23M -) $19M (totals $100M) as they make progress reducing their balance sheet over the year. When you add together the $269M and $31M NII estimates, that yields $299.9M in 1Q26 total revenue. Net interest income is doing a bit of heavy lifting here, but the quarter is also seasonal. This would put total revenue up 1.3% QoQ (1Q25 saw -2.6% QoQ, 1Q24 -8.9% QoQ) and would represent 40.6% YoY growth. The average analyst estimate for 1Q26 today sits at $298.5M, so Upstart is right in line with what's expected in this scenario. For those who prefer the more 'old school' UpstartIQ method, I am also arriving at a rough estimate of $300M for 1Q26 revenue. In this scenario, I am assuming that once again net interest income comes in around $31M (versus $31M in 4Q25), Servicing Platform Revenue of $44.2M (up 3% QoQ - the same growth rate seen in 4Q25) which brings total non-core revenue to $75.2M. With 3,229 1Q26 Trustpilot Reviews and a Review Rate assumption of 0.67% (flat versus 4Q25), I arrive at 481,940 loans originated. Applying a Revenue per loan of $467.90 to that (down 4% QoQ, effectively the same decline experienced from 3Q25 to 4Q25), yields $225.5M in total core revenue. Adding together the $75.2M and $225.5M estimates total $300.7M in 1Q26 revenue, closely aligning with my previous $299.9M estimate. These are not necessarily my final estimates, but it's how I'm thinking about the quarter today. I will be closely monitoring review trends in April leading into ER to try and get a gauge for how 2Q26 is trending. Right now, despite the tumultuous headlines, I am not seeing any direct impact on Upstart visible in 1Q26. Of course, I can be totally wrong on that. Believe it or not, I am not an insider lol. But back in 3Q25, I sounded the alarm when #UTPI fell rapidly, and it turned out Upstart had tightened their models. As of today, I am not seeing anything in the data that is overly concerning. 🫡

English
0
0
16
3.7K
Pat MacNeill أُعيد تغريده
Kevin Bambrough
Kevin Bambrough@BambroughKevin·
Video for $HG $HGRAF education purposes that I made with Google NotebookLM. Its far from a perfect illustration but its pretty good so I thought I would share for feedback.
English
24
46
264
30.1K
Pat MacNeill
Pat MacNeill@MacNeillPatrick·
$RILY Trading down 8% today on the release of the full year results. The main concern might be the adverse opinion stated by the auditors. The Balance sheet has about $500 million in goodwill/intangibles that will need to be valued lower. nasdaq.com/press-release/…
English
0
0
0
364
Pat MacNeill أُعيد تغريده
GFed
GFed@GfedGoCrazy·
April fools doesn’t hit the same living in a misinformation epidemic
English
422
37.1K
273K
3.2M
Pat MacNeill أُعيد تغريده
Culture Explorer
Culture Explorer@CultureExploreX·
According to Edmund Burke, when a civilization carves its story in stone, it isn’t just remembering, it’s teaching every generation what it must live up to.
English
59
526
4.9K
290.9K
Pat MacNeill أُعيد تغريده
Coach Noah Revoy | Arms Dealer For The Soul 🏴‍☠️
I am going to tell you a deep secret of the English language. When we want to tell the truth, we use Germanic-derived words. When we want to lie, we use Latin-derived words. - Germanic (Old English) words → concrete, direct, sensory, testable - Latinate/French words → abstract, bureaucratic, distancing, often euphemistic Death / harm Germanic (plain, testable):kill, die, hurt Latinate (distancing, euphemistic):terminate, expire, neutralize, collateral damage 👉 “We killed civilians” vs “There was collateral damage” Lying / deception Germanic:lie, cheat, hide Latinate:misrepresent, obfuscate, prevaricate 👉 “He lied” vs “He misrepresented the facts” Money / exploitation Germanic:take, steal, pay Latinate:appropriate, extract, leverage, monetize 👉 “They’re taking your money” vs “They’re extracting value” War / violence Germanic:fight, bomb, burn Latinate:engagement, kinetic action, force projection 👉 “We bombed them” vs “We conducted kinetic operations” Bureaucracy / responsibility Germanic:you broke it, you did it Latinate:mistakes were made, systemic failure occurred 👉 Notice how the subject disappears. Money / exploitation Germanic:take, steal, pay Latinate:appropriate, extract, leverage, monetize 👉 “They’re taking your money” vs “They’re extracting value” War / violence Germanic:fight, bomb, burn Latinate:engagement, kinetic action, force projection 👉 “We bombed them” vs “We conducted kinetic operations” Bureaucracy / responsibility Germanic:you broke it, you did it Latinate:mistakes were made, systemic failure occurred 👉 Notice how the subject disappears. If you see a public statement filled with Latin-sounding words, you are being fooled, tricked, manipulated, or lied to.
Tom Rowsell@Tom_Rowsell

English es obviamente un lingua romance al core e would be vastly meliorated by le removal of le barbaric Germanic elements.

English
231
2K
10.3K
832.7K