
ventures @servicenow | opinions my own
somewhere in cyberspace
Joined August 2014
- Tweets196
- Following3.3K
- Followers207
- Likes25.7K
given what we are seeing on cybersec, 5 observations:
1. workload identity will become big in the next few years
2. long lived credentials become irresponsible
3. bullish on behavior detection for workload access at the authorization server (across workloads and for a single workload)
4. using AI to ensure least privilege per task is a must. too easy to find escalations today. scan changes
5. have a way to quickly launch swarms of agents that can act on cybersec to help defense. be ready
N Wang retweeted
🎉 AI Classify + vector search beats frontier models on accuracy and cost!
At Databricks, we see customers with large taxonomy classification problems like vendor name normalization and biomedical entity linking. Typically they reach for regex, trained classifiers, or direct LLM calls, but these methods struggle on cost, maintenance, and context limits.
To address this, we experiment with 3 classification methods:
1. Vector search
2. AI Classify Workflow: shortlist label candidates with vector search, then classify with the Databricks AI Classify function
3. Direct frontier model calls (GPT-5.4 mini, GPT-5.6 Luna, Sonnet 5, Gemini 3.5 Flash) with prompt caching.
We find that the AI Classify workflow yields the best overall accuracy at ~100x cheaper than the next best direct frontier model.
Shoutout to my teammates @arnav_thebigman @ivanzhouyq @nihit_desai !
N Wang retweeted
Octolane started by watching my mom.
Single mom. Four sons. Seven days a week. No CRM. But she remembered every customer and every promise.
I was Mintlify's first intern working in SF. One day I looked at the Salesforce tower and thought: my mom didn't need a CRM. What if no one did?
Dropped out of Duke. Started Octolane with my best friend @halim__rafi and a team of 5, engineers twice our age who left real jobs to bet on us. Working from coffee shop to coffee shop, 7 days a week, spending as little as possible.
Today, after onboarding 1,000+ companies manually, @octolane - The Self-driving AI CRM is generally available.
You talk to Octolane like a co-founder. It figures things out and does the work end to end. Because the best CRM is no CRM.
→ 10,000+ action types - the largest action library in any CRM
→ 200+ integrations + MCP server (works inside Claude, Cursor, ChatGPT)
→ Multi-agent system: agents talk to each other to run your pipeline
→ Self-improving sales playbook that compounds with every deal
Give it a shot today. Break the product. Let us know.
My number: 628-285-1600. Text me for a launch-day coupon. First 20 calls, I'll onboard you myself.
🌉 Built in San Francisco | Thanks to @ycombinator
N Wang retweeted
My lawyers FINALLY gave me permission to spill the beans...
9 months ago we announced that our dorm room startup @styl_app was acquired by an AI research lab run by our "childhood hero". Today, we're proud to reveal that this was none other than @SebastianThrun: the founder of Google X, Waymo, and Udacity.
We founded Styl with the vision that swiping is a great UX primitive to collect data. From February to December 2024, our ~150k users swiped 30M+ times on a dataset of <100k items. This is likely the most dense shopping data in the world, and I'm so excited for Sebastian to disrupt yet another industry.
On a personal front, I've moved back home. The dream is to build a generational Consumer AI company for the world from India, and I'm hacking away with @shruthibadri. We want to build the most high-density talent team the nation has ever seen; if you know engineers or designers we should chat with, please send them our way :)
N Wang retweeted
Just graduated Stanford (🎉), and throughout college I wrote down every single life hack I learned in an Apple Notes doc. Here are my favorites from 4 years of hard-earned lessons:
- In college, try to find a friend, a mentor, and a role model (upperclassman or such). It’s as simple as this.
- During orientation, write down people’s names in a notes docs along with some identifiers for instant uncanny memory. Also you should think of a great fun fact about yourself because you’ll have to deliver it dozens of times
- Get ready for rejection. I was rejected from 25 clubs, classes, jobs, and programs in my first two weeks. It turned out that many of these selective programs didn’t matter much, and I am grateful for the humbling experience.
- Your Stanford (but more generally, every college) email address is powerful. Cold email professors and professionals to set up mentoring relationships. Show up to office hours consistently if there’s a professor you really admire (they’re often surprisingly empty!)
- Get to know upperclassmen to get college specific life hacks
- ON SELECTIVE CLUBS: Many clubs with selective freshman programs can be solid places but ultimately not relevant for real work/company-building in many ways. That is to say, these [redacted] programs are described by alums as being more useful for meeting people than what they actually learned. You are free to make your own path.
- Prevent info overload. Consider setting up your email to auto-mute notifications from mailing lists so you don’t get a vibrating pocket every five seconds.
- Choose your classes, particularly in mid-late stages (soph-junior year), to hit ~3 requirements at once (there’s a technique to this, will make a video). You’d be surprised how quickly you can graduate this way
- Take hard classes with a cracked friend who can explain things. Do NOT use ChatGPT to outline your general approach; you’re kicking yourself in the foot if you don’t learn at least the basic principles yourself
- When the instructors’ lecturers are not working for you (i.e. falling asleep or not really learning), consider not going in person, so you can learn things at your own pace with recorded lectures. I was a stickler for going in person early on, on principle, and realized quite late there was no point to that
- Find a workout buddy and form a routine. Predinner workouts are amazing
- Declare ASAP, even if you have a random advisor. You can always reassign. It’s a good sense of security
- Depending on what you want to do, often it can be a waste of time just to max out your academic difficulty and have no life otherwise. It's not a flex to be constantly sleep deprived and taking an unmanageable load of difficult classes while neglecting to do longterm planning (applying for internships etc)
- At least at Stanford, people can be flaky, i.e. tend to quit things fairly quickly. It often pays to stay on even just a bit longer. For example, many clubs or activities are extremely competitive freshman year, but by the time junior year rolls around you might be the only person left
Disclaimer: These were tips that worked for ME but likely do not work for everyone. Enjoy college, hope this helps!!
N Wang retweeted
🏛️@EffiGov is the Voice AI call center for local governments that answers resident questions, files service requests, and routes calls to the right department.
Only 2% of cities can afford a traditional 311 center - and those that can are still overwhelmed. EffiGov fixes that.
N Wang retweeted
Thanks for making this possible!
Replying to @djkesu1
Winners (presented by @innate_bot ): Raju, the chess bot.
@snbafana
@marco_schonert
@AlexBoniske
@TalhaKhan_TK_
@carrawu
They use stockfish and a trained pick-and-place ACT model to play chess. Bonus: It also smack talks it's opponents!
N Wang retweeted
Try qrsnapr.dag.gy rn
finally solved my problem of having to use my phone for qrcodes
N Wang retweeted
Aligns with my predictions around software development - today's biggest constraint, systems engineering ability, will continue being the biggest constraint to revenue growth.
AI makes it a lot easier to start new projects, but only a little easier build resilient applications.
Some strange side effects:
1/ support tickets have gone up 40% in one month. Builders are emailing us Linux commands that they are trying to run on Postgres 😅.
The @supabase support team have been amazing, with a CSAT of 94.9%
N Wang retweeted
// ALERT //
In honour of Polaris Dawn’s historic milestone of the first commercial spacewalk, Dyson Sphere Maintenance is doing a giveaway.
Like and repost if you’d like a free jacket
N Wang retweeted
Goodbye AIOps: welcome AgentSREs—the next $100B opportunity
"AIOps" is just another buzzword that fails to capture the true potential of AI in IT ops and observability.
We're on the cusp of a fundamental shift in how organizations monitor, debug, and optimize complex software systems.
N Wang retweeted
GLOBAL OUTAGES
- Major banks, media and airlines affected by major IT outage
- Significant disruption to some Microsoft services
- 911 services disrupted in several US states
- Services at London Stock Exchange disrupted
- Sky News is off air
- Reports the issue relates to problem at global cybersecurity firm Crowdstrike
N Wang retweeted
Wowwwww
Everyone in fintech has been saying that the Bilt credit card ($3.1B valuation) is too good to be true; How can you get cash back on rent payments?
Turns out they have an unsustainable deal with Wells that is losing the bank hundreds of $millions. Wells made a number of incorrect assumptions in their underwriting.
They assumed that it would be a top of wallet card for renters who carry balances. Instead, a lot of savvy consumers (myself included) use the bank for the minimum required to collect the rental points.
Wells assumed that 65% of card-purchase volume would be non-rent, generating interchange-fee revenue.
The reality is inverted... most of the volume is rent payments on which they don't earn much.
Wells also assumed that 50-75% of the balances would be revolving, earning them interest. The reality is they mostly have cardholders like me who don't carry a balance - only 15-25% are revolving.
Wells is paying 0.8% to Bilt on rent payments, even though the bank doesn't get interchange on them. Wells has a massive mortgage business and thought that they'd be able to cross-sell mortgages when the renters become homeowners but that hasn't panned out.
Kudos to Bilt for negotiating a killer contract that doesn't expire until 2029... but I wonder what the future looks like after that. Wells says they won't renew it as-is.
In the meantime - it's a great card... pay your rent via the card and make 5 other transactions a month to get 1% back on your rent.
N Wang retweeted
TIL the reason Google Maps doesn’t have a “most scenic” walking route option is is force people to walk through low-income communities
unavailable
now that @matt_levine has a podcast, @BloombergTV should play Money Stuff reruns in their (nonexistent) downtime
1
326
N Wang retweeted
I'm sorry Slack, you're doing fucking WHAT with user DMs, messages, files, etc? I'm positive I'm not reading this correctly.
The most motivating thing is finding places where you learn and think in new and (often exciting) ways. Maximize opportunities to be inspired and pushed into flow states.
2
313






















