
Co-Founder @boomandbucket. Previously Head of Cat Digital Labs and VP of Product at Yard Club. Alum of MIT & CMU.
Silicon Valley
Joined March 2009
- Tweets593
- Following394
- Followers306
- Likes251
Super psyched to welcome Poetiq to YC family— we're over the moon to welcome Shumeet Baluja and welcome back YC W11 alum Ian Fischer.
Getting to the top of ARC-AGI is no small feat, and recursive improvement a powerful milestone.
.@poetiq_ai is a new startup founded by former DeepMind researchers (@itfische & @sbpoetiq) that recently achieved a major jump on the ARC-AGI benchmark by layering a recursive self-improvement system on top of Gemini 3.
In this conversation at NeurIPS, @FrancoisChauba1 sat down with Poetiq co-founder Ian Fischer to find out how they're increasing performance using prompts and system design alone.
They also explore recursive self-improvement, benchmarking progress toward AGI, and why automating prompt engineering may be one of the most powerful levers in AI today.
00:11 — Introducing Poetiq and the ARC-AGI Breakthrough
00:49 — How Big Is the Performance Jump?
01:18 — Ian Fisher’s Background: YC, Google, DeepMind
02:00 — Recursive Self-Improvement Explained
03:00 — Why Poetiq Targeted ARC-AGI
03:58 — Improving Models Without Access to Weights
04:26 — Ensembles, Voting, and System-Level Optimization
05:30 — Why Gemini 3 Changed Everything
06:21 — What’s Next: Benchmarks, Research, and Customers
07:14 — Is Recursive Self-Improvement a Path to AGI?
08:46 — When to Stop Hill-Climbing
09:16 — Automating Prompt Engineers and Agents
Samir Shah retweeted
Sequoia just called the end of an entire go-to-market era and most SaaS companies won’t realize what hit them for 18 months.
Product-led growth was built on one assumption: humans would try the software. The entire playbook since 2010 optimized for human discovery. Beautiful landing pages. Frictionless free trials. Viral invite loops. Slack, Dropbox, Zoom, Calendly. $200B+ in market cap created by winning the user’s first 5 minutes.
None of that matters if an agent is picking the software.
Claude doesn’t care about your hero image. It can’t be impressed by your Dribbble awards. It’s reading documentation, parsing user reviews, checking API reliability, and matching features to use case. All the surface-level polish that convinced lazy humans to click “sign up” becomes irrelevant.
The new PLG funnel isn’t landing page → free trial → activation → conversion.
It’s agent query → documentation scan → feature match → recommendation.
Which means the new moat looks completely different. You don’t need the best onboarding. You need the best documentation. You don’t need viral loops. You need structured data that agents can parse. You don’t need a beautiful UI for the first session. You need an API that an agent can actually call.
The companies that won PLG hired designers and growth hackers. The companies that win agent-led growth will hire technical writers and developer relations engineers.
And here’s the part nobody’s pricing in yet: agents don’t have loyalty. They don’t have switching costs. They’ll recommend Supabase today and something better tomorrow if the documentation is cleaner or the pricing is more transparent. The stickiness that made PLG so powerful, the network effects and learned behavior, doesn’t transfer.
Sequoia is telling you the entire distribution layer is being rewritten. The question is whether your product is optimized for human attention or machine parsing. Most are built for the wrong audience.
Sequoia partner @sonyatweetybird says we're going from the age of product-led growth to the age of agent-led growth.
"You see this most clearly if you're using Claude Code actively. It says, 'Hey, for a database, you should use Supabase. For hosting, use Vercel.' It's choosing for you, the stuff you should be using."
"Product-led growth brought us closer to the vision of 'best product wins,' but ultimately people are still lazy. They can't read all the reviews, and they kind of default to what looks cool on the website."
"Whereas your agent has infinite time to go and make these choices for you. It can go and read all the documentation, read all the user comments, and figure out [what you need] for your use case."
Samir Shah retweeted
The Economist has a great piece on strategy sportsbetting apps use to throttle smart bettors:
▫️Skilled players are “sharps” and given “stake restrictions” if they play too well (bets are capped).
▫️Rest of players called “Square”.
▫️In 2025, 4.3% of active UK accounts had a “stake factor” below the maximum bet allowance of 100%.
▫️Sportsbook will take bets with a profit margin as low as 4.5%.
▫️If they are able to do good “player-profiling” and keep the “sharps” from playing, the profit margin can reach 10-20%.
▫️As important as keeping out “sharps” is hooking “whales”, the deep-pocketed players that are willing to keep playing (and losing) large sums.
▫️Some “whales” are actually “sharps” in disguise, though. They’ll lose a bunch of bets to lull the sportsbook then put down a massive bet when they have an edge.
▫️While there is a risk of a “whale” being a “sharp”, the value of a real “whale” is so high that sportsbook will take the risk
▫️“In March 2024 PointsBet, raised its share of online sports-gambling revenue in New Jersey from 11% to 24% after wooing a single cash-spouting customer away from DraftKings.” (I can confirm that this wasn’t me).
▫️How sportsbook profile players:
> Playing on Mobile is a good sign (where majority of people play)
> Playing on PCs is a bad sign (it’s easier to compare odds and run models)
> E-wallets are a red flag (sportsbooks prefer debit direct deposit that can attach a player to a single account; e-wallet is more anonymized and players can move cash between sportsbook more quickly to shop for the best odds)
> Women bettors are a red flag (most bettors are men and “sharps” often use women to place bets)
▫️First wagers are a major tells (typical bettors go after top leagues — NFL, NBA, EPL — and do so near the start of the game).
▫️Popular bets for “squares”: who will win, scoring margins and how star player will perform (also, they love multi-leg parlays).
▫️“Sharps” go after less popular leagues and place bets as soon as odds are published, when they are most mispriced. They also go after less popular bets such as “pts in Q3” or stats from a random player (“Sharps” rarely do parlays and don’t withdrawal winnings often).
▫️One gambling consultant tells The Economist that “By the time a customer places his first bet, [sportsbooks] are 80-90% certain they know the lifetime value of the account.”
▫️”Sportsbooks look at a player’s ‘closing-line value’ — a measure that compares the odds at which he bets with those available right before a match begins. If it is consistently ahead of the market over his first ten wagers, he is highly likely to beat the book in the long run.”
▫️Sportsbook mathematically monitor players and creates a new risk score every 6-8 hours (risk score = estimate of probability that customers will wind up unprofitable).
▫️E-wallet users, women and bets over $100 are flagged. These suspicious bettors are given 30% of maximum bet (and proven sharps only allowed 1%).
▫️High-skilled players will often get a “beard” to bet on their behalf. Most sportsbooks ban this practice but it is widespread.
▫️Safest “beards” are close friends and relatives because you can mostly rely on them to pay out any winnings. The “beards” try to look like degens (playing at 3am, bet non-stop and doing ridiculous parlays) before placing a winning bet.
▫️The most effective strategy for “sharps” is “whale-flipping”. Find a losing gambler, then ask to put a (likely) large winning bet amongst their pool of guaranteed losers.
▫️Once “sharps” max out the people they can use as “beards”, they tap professional networks called “movers”. These “movers” employ a bunch of “mules” who can put down bets on the behalf of the network. Low-end movers charge 10-20% while high-end movers charge 50% of winnings.
***
Lots other great details here: economist.com/christmas-spec…
Samir Shah retweeted
A massive new study on peak performance included 34,000 international top performers: Nobel laureates, renowned classical music composers, Olympic champs, and the world’s best chess players. It shows early specialization is a trap, and the road to greatness is long and varied.
Samir Shah retweeted
Is more intelligence always more expensive? Not necessarily.
Introducing Poetiq. We’ve established a new SOTA and Pareto frontier on @arcprize using Gemini 3 and GPT-5.1.
Samir Shah retweeted
Be like Amazon.
> secretly build an AI coding assistant
> save $260M in costs and eliminate 4,500 years of developer work
> now quietly sell it to all the Fortune 500s
I got curious about their entire business model and dug deeper.
What I found:
Samir Shah retweeted
NEWS: Billionaire Michael Bloomberg has released a new opinion piece titled “NASA's $100 Billion Moon Mission Is Going Nowhere”
“The more I have learned about Artemis, the more it has become apparent that it is a colossal waste of taxpayer money. The problems start with the mission, which is more political than science.”
Full opinion piece: bloomberg.com/opinion/articl…
Samir Shah retweeted
𝗘𝘅𝗰𝗶𝘁𝗶𝗻𝗴 𝗡𝗲𝘄𝘀 𝗳𝗿𝗼𝗺 𝗕𝗼𝗼𝗺 𝗮𝗻𝗱 𝗕𝘂𝗰𝗸𝗲𝘁!
We’re thrilled to announce our partnership with @procoretech. This collaboration launches a groundbreaking initiative to standardize construction equipment data, significantly enhancing efficiency across the construction industry.
Discover more about this innovative initiative and read our detailed press release here: buff.ly/4eRI2dv
#ConstructionTech #Innovation #Partnership #Procore #BoomAndBucket
Samir Shah retweeted
Don't miss out—these deals won't stick around! ⏳
Time is running out to score these last chance heavy equipment offers. Check out this link to explore what's still available → buff.ly/3ZTxyFx
What machine would be a game-changer for you? Drop your thoughts below! 👇
Samir Shah retweeted
0/ 80% of 100 dollar bills are held offshore, representing over $1.5T in value!
For most, outside of the global elite, the only access to USD was cold hard cash.
Stablecoins introduced digital dollars to the rest of the worlds, and the adoption metrics are astounding. 👇
Samir Shah retweeted
I was talking to an Indian American founder yesterday. He mentioned that after exit, his biggest lifestyle change is that he doesn’t need to do laundry or vacuum + has a full time chef and trainer.
Then he said - “It’s like my life in India”
Samir Shah retweeted
You absolutely cannot imagine the amount of carbs and sugars that Indian people eat daily 😅
Samir Shah retweeted
1/ It's easy to get lost in the euphoria leading to @Blast_L2's June 26 TGE, but Blast's embrace of next-level developer incentives will be it's lasting legacy.
Let's explore this revolutionary co-marketing program known as Blast Gold and learn why it's exciting! 👇
Samir Shah retweeted
David Attenborough is now narrating my life
Here's a GPT-4-vision + @elevenlabs python script so you can star in your own Planet Earth:
Samir Shah retweeted
The first 5 months of Boom and Bucket’s history were spent cold calling people who were selling heavy equipment on Craigslist.
Me and my cofounders @klinear and @sshah1983 probably called 500 people.
Why?
Because we were still vetting the opportunity.
We wanted to validate a few things.
a) if our idea actually sucks or not
b) if we like or dislike serving the customer
c) if we had any blindspots we were unaware of
Basically, we were trying to find reasons NOT to commit to this idea.
This comes back to something I believe…
There’s a huge asymmetry in making commitments.
It takes 5 seconds to decide to commit to something.
But it takes 5-10 years to actually show up every day.
If there aren’t at least a few months of idea validation prior to making a commitment, you’re screwed.
It’s much better to “waste” 3-4 months of time to find out that your idea wasn’t as great as you thought, than spend 1-2 years finding out the hard way.
It’s just a waste of time.
Your job in idea validation is to find out as many of the bad things as possible up front.
And the process of finding out the bad stuff isn’t fancy.
At the end of the day it’s just talking to a bunch of people.
- Cold call who your customers would be
- Talk to people who invested in the space before
- Go talk to previously failed entrepreneurs in the industry, why did they fail? What are their thoughts today? Is your strategy different?
If you’re pursuing a new idea, you have to balance excitement vs. conviction.
You need to turn over enough stones to make you LESS excited and MORE convicted.
Otherwise, you’re wasting your time.
Samir Shah retweeted
I’m Vegas bound for CONEXPO-CON/AGG and @AGCofA
If you’re there, I’d love to introduce you to @boomandbucket . I should be easy to find as I look just like my profile picture with mustache and cowboy hat.
Oh, and we’re doing a sweepstakes for ONE MILLION BUCKS! I hope you win! Enter here: lnkd.in/gdmjnt9u
#contractors #agc #conexpo2023 #heavyequipment
Samir Shah retweeted
We are glad to share with you the contractor of this week: @bcmasongroup 🤝
⠀
Thank you for your trust and for sharing your experiences with us. 👏
#contractoroftheweek #boomandbucket #constructionequipment #heavyequipmentlife #heavyequipmentforsale































