Joined November 2022
- Tweets 1,307
- Following 816
- Followers 1,310
- Likes 2,210
209 Photos and videos
Pinned Tweet
25 Sep 2025
We raised $4.1M for this moment: introducing Structify Flows.
It's the first “vibe data engineering” platform that lets you do enterprise-grade data work in minutes:
With Structify, just describe what you want in plain English.
No code. No hiring. No backlogs.
119
68
772
1,103,726
Palantir has kept these tools in an isolated, walled garden that price excludes the midmarket and refuses to allow the enterprise market to self manage.
If you want that era to end, try structify. DMs open.
Jun 4
“Pointing an LLM at hundreds of disconnected, ungoverned databases gets you a system that hallucinates, is insecure, and unauditable. For something as consequential as our nation’s agricultural data, that is not just useless — it’s dangerous. The Ontology has been the key to delivering AI-enabled technology to every farmer in the country.”
At AIPCon 10, the USDA demonstrates how the Ontology now underpins national food supply security.
1
1
3
207
Alex Reichenbach retweeted
Jun 4
Let’s go Knicks 😂
Great seeing @alexrkonrad @shensi @bernhardsson and @ryanjdaniels today.
General consensus: NYC > SF on all fronts
1
2
22
1,687
Great seeing @alexrkonrad @shensi @bernhardsson and @ryanjdaniels today.
General consensus: NYC > SF on all fronts
1
1
20
2,923
My favorite structify workflow. It pulls from all my outreach methods and my CRM, and lets me know who I need to reach out to by slacking it to me.
3
217
Very much agree with this. It is so easy to build a bad product now and still hard to build a good one.
People just have a temporary illusion that their internal teams will love using their badly built internal tools.
Internal AI tools will be built in weeks, then abandoned in months.
That's not a build-vs-buy debate — it's a maintenance reality most teams haven't confronted yet.
1
6
139
To be fair. Copilot responses do make me laugh
Apr 5
Copilot is ‘for entertainment purposes only,’ according to Microsoft’s terms of service techcrunch.com/2026/04/05/co…
1
184
We’re on it! We let the semantic layer get defined through chats. And a native claude plugin is in beta - feel free to dm.
Apr 5
This may be a dumb question but I’ll ask it here anyways:
I can’t find a good way for my various AI chats to automatically sync its conversation history into a structured knowledge base. So that as I update various chats from time to time and refine context, my knowledge base automatically grows with this new info.
118
It’s unfortunate but required. Vibe coding is a security risk that apple doesn’t want.
Let’s stop vibe coding nonsense and actually figure out how to make AI work at work.
Mar 31
Apple's crackdown on vibe coding apps puts it on the wrong side of history
vibe coding is the biggest democratization of software since the App Store itself. But Apple is blocking the tools that make it possible.... the same week it turns 50.
The irony is that Apple should be leading this moment and instead its holding it back
w/ @jaswu_
2
2
146
We’ve been tackling this at Structify recently. Doing practical, responsible AI is both annoying and important. RBAC on every context node. Every data source needs different permissioning. Every query needs different permissioning.
It might be annoying, but if you don’t do it, you have a countdown until your next data leak.
3
194
Very much agree. We still pay for Hubspot. I have no interest in trying to reinvent that wheel and find all the edge cases from scratch.
Apr 4
SaaS is dead.
Why pay $45/month when Claude Code can build you anything for $200/month?
Except there's an issue.....
Most people don’t want to spend their evenings or work ours prompting AI, debugging edge cases, and figuring out why the payment process is broken again.
I'll die on this hill, but people have always paid for convenience.
They want to open an app and have it work.
The best apps look simple.
The backend is anything but.
That gap — between what users see and what actually runs underneath — is where all the work lives.
SaaS is dead. Of course, for those who want to engage in bait.
100
It's this type of stuff that makes me think local dev environments shouldn't be trusted anymore.
Apr 2
URGENT PSA - New supply chain attack vector that I found WILD > AI LLMs hallucinate package names roughly 18-21% of the time.
Hackers have started pre-registering those hallucinated names on PyPI and npm with malicious payloads; they call it "slopsquatting"
You can only imagine what's next
Community note
The 'slopsquatting' attack vector was documented as early as April 2025 and not newly discovered. The cited 18-21% package hallucination rate applies to open-source LLMs; commercial models average 5.2% according to the referenced study using pre-2025 models. socket.dev/blog/slopsquat… arxiv.org/pdf/2406.10279
1
198
Alex Reichenbach retweeted
Apr 2
URGENT PSA - New supply chain attack vector that I found WILD > AI LLMs hallucinate package names roughly 18-21% of the time.
Hackers have started pre-registering those hallucinated names on PyPI and npm with malicious payloads; they call it "slopsquatting"
You can only imagine what's next
Community note
The 'slopsquatting' attack vector was documented as early as April 2025 and not newly discovered. The cited 18-21% package hallucination rate applies to open-source LLMs; commercial models average 5.2% according to the referenced study using pre-2025 models. socket.dev/blog/slopsquat… arxiv.org/pdf/2406.10279
65
191
1,594
574,846
Alex Reichenbach retweeted
Apr 2
does your data look like a frat basement floor? call 929-460-9925
1
2
192
Mar 31
I'm done upgrading any dependencies. No dependencies in 2026 can be trusted.
87
Mar 25
Here's what actually happens.
Three executives at one of our clients presented three different NRR numbers at their sales kickoff.
All three used AI. All three thought they were right.
Guess who had to clean it up? RevOps.
Until now.
1
5
108
Mar 25
The problem was never writing queries or building dashboards.
Any LLM can do that now.
The real problem is knowing which table to pull, what "active user" actually means at your company, and why revenue lives in four systems with four definitions.
That's not a data problem. It's a context problem.
5
82
Mar 20
I just watched an agent not want to do my task and assign it to a subagent instead.
AGI has arrived
1
1
163