Simrat retweeted
Working with agents all day is more tiring than writing code ever was, and we must admit it.
Typing wasn't that tiring, you knew what you wanted, your hands did it, your head and thoughts wandered a bit.
My days now are 3-4 agents running all the time, and each one stops every few minutes with a question. Approve this command? Option A or B? This test fails, should I change the test? Should I continuue?
Each one is a an actual small decision, but each one needs the full context of a codebase and new code I didn't write line by line.
By 6PM I'm done, not because I wrote a lot, but because I decided a lot. For a while I thought my limit was tokens, then the weekly usage limit, but turns out it wasn't. My limit is how many good decisions I can make before I start approving everything just to make the questions stop.
How many agents do you run in parallel and for how long before your own answers get worse?
Simrat retweeted
Unity $U +23%
Posted its first GAAP profit since its IPO
Expects its ad business to more than double next quarter
Grew revenue 32% while cutting sales & marketing spending by 38%
Added $587 million in cash this year
UNITY $U Q3’26 EARNINGS HIGHLIGHTS
🔹 Revenue: $622M (Est. $567M) 🟢; +32% YoY
🔹 EPS: $0.25 (Est. $0.05) 🟢
🔹 Net Income: $122M
🔹 Adjusted EBITDA: $257M
Q4 Guide:
🔹 Strategic Revenue: $683M-$691M
🔹 Strategic Grow Revenue: $515M-$520M
🔹 Strategic Create Revenue: $168M-$171M
🔹 Non-Strategic Revenue: $5M
🔹 Adjusted EBITDA: $290M-$295M
Segment Net Revenue:
🔹 Grow Solutions: $457M; +44% YoY
🔹 Create Solutions: $165M; +8% YoY
Announcements:
🔹 Vector: +126% YoY; +40% QoQ
Other Q3 Metrics:
🔹 Free Cash Flow: $220M
🔹 Operating Cash Flow: $224M
Comments:
🔸 “Third-quarter performance was our strongest ever, with Vector growing 126% year-over-year and 40% quarter-over-quarter, and the first GAAP profitable quarter since our IPO.”
I don’t recall when was the last time that I had 2-3x more ideas than my 20 stocks PF.
I believe things are truly at cusp of inflection and maybe valuations somewhat reflect that. Second, Discovery is becoming easier. Lets see what transcends from development to commercial stage because failure rates might remain same.
How reading global concalls help -
1. Reading that bearings are in supply crunch led to me buy bearing rings company
2. Reading Spotify creating friction for people on free tier in India led me to buy music label in India
3. Reading castings and forgings are also in supply chain crunch helped me to allocate more to existing forgings/castings co in portfolio.
A power ancillary which is quality company is doing 700 Cr QIP to fund WC and Capex. The demand is very strong for them, but then why do QIP, balance sheet is already clean, take loans for this and pay them back as demand is strong. Why dilute shareholders permanently. And then people say corporate balance sheets are cleanest than ever.
Only in India, no one questions QIPs until stock is doing well, in US by now it would be 20-30% lower circuit.
Btw, same company is thanking employees too much in recent concall, hope everything is fine internally and execution issues don't prop up later.
In bulk commodities, 99.5 percent purity is identical regardless of who makes it. In aroma chemicals, the remaining 0.5 percent trace impurity profile completely changes how a molecule smells to a master perfumer’s nose.
Every fragrance house approves a vendor not just on chemical specification, but on "olfactive acceptance."
It is fascinating to read about period during and before Middle Ages. People used to die so quickly due to so much of random stuff.
In 1300s - A Black Death event similar happened where 200 million out of 500 million people of the world died. 40% of population of world gone. We are all the product of some ancestors who probably were less social or had much better immune systems to survive the bacteria of this event.
Humanity has made so much progress and we forget how bad living conditions were even 100 years back.
It is interesting to note that likes of Netflix have structural headwinds of shortening attention span while a music streaming platform probably does not have the same headwind.
But maybe both can have structural tailwind of people getting rid of mundane tasks that AI would be doing instead (only if jobs remain :D).
The story might have legs:
1. There is multi phase expansion in place which would need time to get audited for more CDMO work
2. Existing CDMO customer should do sustained business. A global pharma innovator does not spend months auditing a factory, transferring proprietary process chemistry, and validating batches just to buy for four quarters and walk away.
FY26 Baseline: ~500 to 535 kiloliters (KL).
Phase 1 (Q2 FY27 - Completed Ahead of Schedule): Commissioned 614 KL in Q2 FY27 (beating the initial 600 KL target).
Phase 2 (Target Q2 FY28): Civil work has already started to take capacity to 800 KL by Q2 FY28.
Phase 3 (Target FY29): Expansion to 1,000 KL.
Phase 4 (Target FY30): Expansion to 1,200 KL, more than doubling the company’s FY26 manufacturing base.
youtu.be/sS5Ik4HCW7Q?si=fhBu…
Morepen interview
Two major CEOs contesting on X. Peak competition
release all the numbers.
does 80% yoy become 65% when you remove passthru inference?
what does it become when you remove ai natives?
and how much of it is from one time software license --> service transition?
fcf, RPO, cRPO?
op margin non-gaap, gaap? dilution?
curious ppls want to know
ps: @alighodsi happy for your success. the world is big and life long.
teapaw.org/bhorowitz/status/21150…
Lot of businesses with below mental work end up working best:
Supply a relatively small fraction of the end product's cost, yet purity at parts-per-billion levels, or tolerance levels for a deviation are low aka multi-step reliability is mission-critical.
Gemini 4 on Antigravity is different beast
🚨Gemini 4 Argon @GoogleDeepMind with @antigravity
has made a significant leap on our RSI-Inference Bench rsi-inference.com with a close tie with GPT-6-Astra.
It's the first model outside @AnthropicAI & @OpenAI to pass a Day-0 Model Support task ✨
Congrats @GoogleDeepMind @PichaiSunder 👏!
Simrat retweeted
🚨Gemini 4 Argon @GoogleDeepMind with @antigravity
has made a significant leap on our RSI-Inference Bench rsi-inference.com with a close tie with GPT-6-Astra.
It's the first model outside @AnthropicAI & @OpenAI to pass a Day-0 Model Support task ✨
Congrats @GoogleDeepMind @PichaiSunder 👏!
We just added Google's Gemini-4-Argon to RSI-Inference Bench 🎉
Lands at #6 of 14 on the joint leaderboard, right behind GPT-6-Astra.
•Day-0 support: 6th, 81% of hidden tests passed (Pass@1 6.7%)
•Inference optimization: 4th, 1.39× speed-up
