swyx
@swyx
achieve ambition with intentionality, intensity, integrity & insanity.
affiliations:
- @smol_ai
- @dxtipshq
- @cognition
- @aidotengineer
- @latentspacepod
3.9K Following 180.8K Followers
worktrees must die
this is 20GB of repeated node_modules lol
among ai leaders i seem to be in the minority in that i am STILL actively using /loop and /goal....
... and i think all of u guys who stopped using it are wrong - not wrong forever, just giving up on it too early in the g5.6/c5 era
now you use it when:
1) you want the right mix of steerability and autonomy
2) you want open ended "loop that generates loops" type end state without deeply specifying path to get there
heres an example of when having a goal on saved my ass in a very long action-reasonining turn
Show more
Yesterday I cohosted a dinner with @dexhorthy with a wonderful group of founders, to talk about agent loops and loop engineering.
Some interesting insights:
* Most of our group was *not* actively using /loop in Codex/Claude Code
* You can build long-running autonomous agent loops through multi-agent handoffs, event triggers, or….just stacks of cron jobs (?)
* Almost everyone believes that no one will be reviewing code in 1-2 years.
* The more interesting question is whether we’d be reviewing *anything* in 1-2 years.
* AI is still a bit of a skill issue. Humans are responsible for maximizing AI output and reducing slopification.
* Will human intellect provide alpha as models get better, or will the playing field be leveled? Most people think it will be leveled a bit, but there is a need for humans to provide alignment, guardrails, judgment, creativity.
* The minimum amount of context you need for AI could be just the codebase with some documentation. Any research/plan files are for one-off tasks and not meant to be maintained. Having a self-organizing wiki is nice but adds complexity.
If you missed this one, we’ll be hosting more dinners like this on a regular cadence! If you have thoughts on what we should talk about, let us know :) (e.g. continual learning, RL envs, competitive differentiation vs. Anthropic, etc.)
Show more
did some yapping about how we’re approaching continual learning :)
noticed that the perjorative connotation around "vibe coding" has completely disappeared since ~everyone, from nontechnical to supertechnical, is now doing it
Slop is code you dont read. And better models means more slop.
the solution? write sloppy tools to build stable systems.
but for the most committed people, go build new foundational systems. Rebuild github, rebuild tmux, rebuild database, and yes rebuild programming languages.
Big thanks to @swyx and the whole community for having us out at @aiDotEngineer
Show more
this, by the way, is the correct way to kill US proliferation of chinese open source
btw as a huge music person i moved from Spotify to YouTube Music 2 years ago and never looked back
1 - comes with youtube premium
2 - much friendlier to both big and small artists
3 - MUCH bigger selection of indie/long tail/concert/acoustic/podcast music
4 - ui doesnt piss me off every single time i open the app
bonus:
- didnt have a failed takeover attempt putting paywalls on the open podcast ecosystem
- doesnt have overfamiliar ai dj telling u dumb things you already knew
cons:
- spotify wrapped is better than youtube wrapped
Show more
I have trouble reconciling how
1) Spotify used to have a very strong eng culture, huge on quality.
2) Their products are extremely buggy, unreliable, and getting worse esp this year. It got so bad I had to offboard from their video podcasts product, just did not work. Here:
Show more
@ArtificialAnlys ok @openai gets it
We are committed to pushing the model frontier across cost efficiency, capability, and speed.
Starting today, we are reducing prices for GPT-5.6 Luna by 80% and GPT-5.6 Terra by 20% , and offering a faster option for GPT-5.6 Sol in the API.
Luna and Terra’s lower prices are reflected in how usage is counted in Codex and ChatGPT Work, so your usage goes further.
Show more
AI engineers are rediscovering ontologies as a way to keep probabilistic agents inside deterministic boundaries. We delve into presentations from @coyle_frankp and @emileifrem from the @aiDotEngineer World's Fair, to see how technologies from the old web are being used to keep LLMs honest and agents on track.
Show more
🆕 First Steps Toward Automated AI Research
Humanity advances by trying things, finding the shortcomings, and fixing them. @RichardSocher keynotes our first-ever Autoresearch track to show how @Recursive_SI is building a Eureka machine for recursive self improvement... and how far we have yet to go.
Show more
today i tried to sign up for @docusign and was presented a captcha that i could not solve
this is an $11B company that everyone hates
people who say "make something people want" dont have the full picture
Show more
Idea: Business owners should crowdsource a list of Most Hated Software and then indiehackers should pick thru and make new clones of them are just "simple" - rewind 10 years of enshittification on them.
I hate (and use):
- dropbox
- gusto
- zoom
- loom
- canva
- accel
- most of gsuite
- substack
- descript
- youtube
Show more
Agents are handling more and more of YOUR money — so much so that AIE NYC will be our first Finance-mainstage conference this Oct!
🆕Live now: the entire WF26 AI in Finance Track!
Watch ↓
@brendanh0gan, Morgan Stanley
Lucas Palma, Nubank
@YogiNotTheBear, FactSet
@Saikallis9012, Fidelity Investments
@menkesu, Intuit
Divakar Kumar, FlyersSoft
Shawn Chan, China Resources Holdings
@siddanth2486, Auditoria AI
@vinooganesh, Kepler
@ShreyaR + @aman2304, Nubank (with Snowglobe)
Show more
I gave a talk on forward deployed engineering to a thousand AI engineers at @aiDotEngineer World's Fair.
A year ago I'd have opened by explaining what FDE stood for. Not this time.
Thank you to @swyx, @realbasilchatha and the AIE team!
writeup / video:
Show more
One of my most defining experiences was 15 years ago watching two ppl with roughly same followers. One HAD to tweet the zinger, HAD to argue with everyone. The other mostly stayed positive and built.
The former ended up having a middling career. The latter a monster career.
Show more
re: hiring right now
it's a huge bull market for AI-native IC's/player-coaches
it's a huge bear market for "heads of X" managers
never seen such furious bifurcation. to oversimplify: 1 year experience managing 10 agents > 10 years experience managing 10-100 people
Show more
incidentally, $ per input/output tokens died as a relevant cost measure sometime last year
if you haven’t updated your x axes to $/task per @ArtificialAnlys then idk if you can be taken seriously anymore these days
Show more
one thing i think people dont appreciate enough about @poolsideai is their unusual degree of openness — not only have they shipped an excellent Small model that somehow beat @thinkymachines at coding, but most people (like @eliebakouch) have been shouting out their excellent papers, but also they're among a rare few to actually expose their full eval dataset as well - beautifully published, with across 6 public benchmarks with 4 runs each and hundreds of turns per run. you can satisfy for yourself if they rewardhack. brilliant.
Show more
in case anyone was wondering, 300 or so people were referred by chatgpt this year
This pod was an incredible gift to the community:
not only our first pod about @xAI, but Ethan really indulged on all our questions on how to train a SOTA Videogen world model, including specific areas (consistent extending/editing, voice) that Grok @Imagine is *still* SOTA,
on top of the factual overviews he ALSO came loaded with opinions/predictions:
- why he's quitting Videogen for LLMs: video models get most of their intelligence from LLMs, not from scaling video data
- why the next frontier for videogen also happens to be video agent models - agentic models trained to orchestrate video models
- why deterministic compression (like MP4) is a useless target vs VAE compression
- Videomaxxing: if you truly believe in the "Moore's law" of AI/genmedia, then video models become the final boss UI of everything, like Flipbook (below)
Show more
this seems quite doable in the space of a single 2-3 hour workshop — any brave soul want to try to livecode this for people as a learning exercise?
