CTO at Cadence - linkedin.com/in/erichauser/
Joined June 2008
- Tweets 2,457
- Following 1,384
- Followers 415
- Likes 1,241
27 Photos and videos
May 27
We have to hit a 40B ARR run rate to have product-market fit now?!?!
May 27
Given the recent burst of activity around enterprise pricing and contracts, I think April 2026 was the month when both OpenAI and Anthropic found product-market fit simonwillison.net/2026/May/2…
3
2
129
May 16
Forked an experiment for an ephemeral Postgres for faster unit testing (github.com/ewhauser/postgres). 2.5x faster with no changes; 3.3x with minor schema setup changes. Day's worth of background work; imagine someone who knows PG internals could get a much larger improvement
1
1
3
161
May 6
The under-discussed shift in AI pricing is that flat fee coding subscriptions are mostly a temporary packaging illusion
1
46
May 6
Enterprise users have already been moved toward seat fees plus usage-based pricing. My guess is post-IPO for OpenAI/Anthropic you will see the flat monthly fee tiers significantly throttled compared to today - if not gone completely.
31
Eric Hauser retweeted
May 1
you know what
all of these "which is better" polls are silly
use codex or claude code, whatever works best for you
i am grateful we live in a time with such amazing tools, and grateful there is a choice
2,176
1,102
22,934
1,642,215
Apr 30
Excited to open up shuck, which is a modern replacement for shellcheck written in Rust - news.ycombinator.com/item?id…. shellcheck is a fantastic project but is extremely slow and hard to modify due to it's licensing. shuck is a clean room reimplementation but should mostly be a drop in replacement. Excited for people to try it out
1
1
113
Apr 22
The one major thing that holds Go back from being the best language for agents is nil. Agents write Go code that is so much more defensive than a human because a human protects from nil upstream while agents don't have that context. Makes agent Go code harder to read.
Apr 20
I'm writing Go again (for what, you'll see later...). `go doc` and `gopls` are like agent superpowers and its shocking how productive agents are out of the box at writing [good] Go code versus other languages I've used (including the JS ecosystem). Also, Go Zig is a good mix.
Go for the higher level and concurrent stuff and then no-libc Zig code plus the Zig compiler for zero dependency cross-compiled cgo with high-performance characteristics (minimize cgo boundary crosses). Chefs kiss.
Its funny because a lot of the shitty ergonomics of Go CLIs like `go doc` and `gopls` (prev. stuff like `go oracle` or `guru`) are totally obviated by agents and not just that but in a twist of irony they're excellent for agents.
Don't worry, its not Ghostty. Ghostty and libghostty will remain pure Zig; it's a fantastic fit and a perfect pairing. This is for something else.
"Wait, I thought you said Go has no place anymore?" I was wrong, mostly because agents are so productive at Go. I won't bring in other languages in this discussion because I don't want to feed the crabs, so to speak. lol.
1
81
Apr 22
Agree with the general sentiment though. We have been all-in on Go and agentic coding since the days we were cut and pasting from ChatGPT. Go and coding agents are incredibly productive but the best language for agents won't have nil.
33
Mar 24
Next step for gbash is to get a working Python implementation. Models really want to use Python when they have a bash environment. Trying two different experiments here - a cgo approach using Monty github.com/ewhauser/gomonty and a pure Go approach using starlark-go
1
1
87
Mar 24
So far the Monty experiment seems to be working well and low friction as long as you don't care about the cgo aspect cc @pydanti_ai
27
Mar 24
10 years ago, you could take an Uber pool from the Marina to SOMA for $6 because the rides were VC subsidized but now it's $20-30. $20 coding subscriptions will be the new $6 Uber rides when the AI companies go public.
1
3
381
Mar 19
So cpython is going to be rewritten in Rust huh?
Mar 19
We've entered into an agreement to join OpenAI as part of the Codex team.
I'm incredibly proud of the work we've done so far, incredibly grateful to everyone that's supported us, and incredibly excited to keep building tools that make programming feel different.
1
70
Mar 19
Codex app is so good but we really need a “PR mode”. 1) Commit, push, and PR at the end of a run 2) prefix the thread name with PR # 3) Loop on the PR until CI passes 4) Address review comments until Codex review give thumbs up 5) Change dot to green cc @thsottiaux @embirico
1
39
Mar 16
GitHub really needs to add automatic merge conflict resolution with AI. That is a feature I would definitely pay for.
1
61
Mar 16
Like I don't even want to see that there was a conflict. Just resolve it, update my branches
26
Mar 16
OS level sandboxes are the end game for AI agents but virtual bash envs - pioneered by @vercel just-bash - work for many use cases. I’ve been hacking on an implementation in Go over the past couple of weeks. Not ready for prime time yet but sharing early: ewhauser.github.io/gbash/
1
1
182
Feb 28
Everyone is this thread is overthinking it. AI writing is easy to spot due to proper use of the English language
Feb 27
My current smells of AI slop/writing:
1. Use of em dash "—". I haven't seen anyone seriously use this over a hyphen "-". Double points for wrapping — or making double points in a single sentence — give it away
2. Making a statement and then colon: like this
3. More subtle, but tone that is off-character for the authors
123
Feb 7
People seem to forget that Heroku has been owned by Salesforce for 15 years. The Heroku you remember was long gone by now
3
162
Jan 23
As usual, @martin_casado is on point here. The exception here is engineering work like technical debt, performance, etc. If the agent can verify it's own work, then the productivity gains have been exponential.
Jan 22
I work with multiple companies where nearly all code is AI generated now. However, the productivity probably has only increased 20-30%. Why?
I suspect because writing code is really running code. Changes are the result of a business learnings. Or an operational learnings. For mature companies, the majority of PRs are sub 10 lines codifying these learnings.
AI clearly helps here (e.g. debugging, running tests, building tools) but less so. Operations and business learnings are workload and company specific.
Until AI can perfectly predict what the market needs, or how a system will be used this bottleneck will exist.
2
285