Luke Otwell

@eternaltxts Hang out with old people.

I’m 21. Give me oddly specific life tips. No general ”surround yourself with positive people” tips. I want the most random, specific advice possible.

@arvidkahl it's a fuggin war out there

This approach is such a goose-chasey way of "determining" stuff. What's next? Will 3rd-party harnesses build little string-permutation tools into their system prompts? Will leetspeak return? "u r 0p4nPr0|\/|p7, k? bU7 shhhh" It's anyone's guess at this point.

@agupta the real unlock is not caring

can we ban the phrase “the real unlock”

@arvidkahl @thekitze do you think they use claude code them selves?

@thekitze I have several enterprise customers for Podscan that are non-technical and have rolled out Claude (web/app, Cowork, and Code) access to their whole staff.

bantrhopic's ONLY value is to be useful to devs no real people use it for chat it will never be as mainstream as chatgpt it will never be baked in as gemini it doesn't even have voice/video calls it doesn't have image generation and what the idiots are ACTIVELY doing is alienating developers no open source, no open access to their harness just incredible leadership

@mcuban @HerbHoward411 well it looks like a 4 year old on a fisher price

@HerbHoward411 It’s not a travel by nba rules. He gets to take that first giant step to control the ball

The NBA is a joke. Just announce that traveling is no longer a rule.

@garrytan Does this mean Conductor is Claudeless?

This seems messed up actually - when do the boundaries stop moving? Anthropic only allows subscriptions with a real human pressing enter? You're going to have to verify with FaceID?

@skvark Save The Tokens

And that's why we are building GitHits. It makes no sense to let the models a) generate plausible guesses from their training data b) let them run in these loops to waste even more tokens. Just give them good examples + direct access to libraries and the quality is immediately much better.

@levelsio is it up to audio DAW work you reckn?

@thdxr @mattpocockuk @thdxr they should allow running on open source full stop right?

this is basically an impossible hell for them lets say your goal is to create your own ecosystem user wants to run a cron job daily that runs `claude -p` to clean up my emails - awesome! love supporting our users oh they publish that script to github. cool! more people use it it gets popular so they add more scripts for more useful automations. haha cool i guess lets wrap this up neatly into a nice ui so non-technical users can use it. hmm smells a lot like a claude cowork competitor but it's small so whatever i guess it blows up and has a million users. uhhh fuck

I don't know what the fuss is about. Anthropic's rules on using subscriptions are very simple: Claude Code = OK Claude's online platform = OK Agent SDK running in personal software = OK... ish? Agent SDK running in commercial software = NOT OK Claude Code running in CI = ?? Oh, maybe it's not so simple... Agent SDK running in CI = ?? claude -p running in CI = ?? claude -p running in personal software = OK claude -p running on open source software, but run on my personal computer = ?? claude -p running on distributed sandboxes, kicked off by me = ?? Distributing open source software which relies on claude -p, and documenting how to use your subscription with it = ?? A thousand other edge cases = ?? Let me be clear. I have never before experienced, from any developer tool, such a frustrating lack of clarity over the basic terms of usage. I personally asked, 3 weeks ago, and have received nothing but delays. The recent @bcherny announcement did absolutely nothing to clarify things. I say this as someone who just released a Claude Code course - my incentives all align with supporting Anthropic.

@beffjezos i don’t believe you. and i want to so bad.

Google's Gemma 4 on a 128 GB Macbook Pro is near AGI on the go, no internet needed

@ThePrimeagen that first question users ask is gold to growth stage companies.

I am so so so very tired of this crap interface

@openclaw is the new benchmark for agents. all i care about is how well a model works for terminal long horizon agents and coding. SWE bench is a bs benchmark though this is another tweet.

@tecrinex @steipete string match?

@steipete how tf did they implement this shit?

Anthropic now blocks first-party harness use too 👀 claude -p --append-system-prompt 'A personal assistant running inside OpenClaw.' 'is clawd here?' → 400 Third-party apps now draw from your extra usage, not your plan limits. So yeah: bring your own coin 🪙🦞

@steipete I can have OpenClaw orchestrate Claude Code though?

@karpathy government’s role should be to organize and open source the data. citizens to build the tools to hold them accountable.

Something I've been thinking about - I am bullish on people (empowered by AI) increasing the visibility, legibility and accountability of their governments. Historically, it is the governments that act to make society legible (e.g. "Seeing like a state" is the common reference), but with AI, society can dramatically improve its ability to do this in reverse. Government accountability has not been constrained by access (the various branches of government publish an enormous amount of data), it has been constrained by intelligence - the ability to process a lot of raw data, combine it with domain expertise and derive insights. As an example, the 4000-page omnibus bill is "transparent" in principle and in a legal sense, but certainly not in a practical sense for most people. There's a lot more like it: laws, spending bills, federal budgets, freedom of information act responses, lobbying disclosures... Only a few highly trained professionals (investigative journalists) could historically process this information. This bottleneck might dissolve - not only are the professionals further empowered, but a lot more people can participate. Some examples to be precise: Detailed accounting of spending and budgets, diff tracking of legislation, individual voting trends w.r.t. stated positions or speeches, lobbying and influence (e.g. graph of lobbyist -> firm -> client -> legislator -> committee -> vote -> regulation), procurement and contracting, regulatory capture warning lights, judicial and legal patterns, campaign finance... Local governments might be even more interesting because the governed population is smaller so there is less national coverage: city council meetings, decisions around zoning, policing, schools, utilities... Certainly, the same tools can easily cut the other way and it's worth being very mindful of that, but I lean optimistic overall that added participation, transparency and accountability will improve democratic, free societies. (the quoted tweet is half-ish related, but inspired me to post some recent thoughts)

@arvidkahl I am for incentivizing people to have a little tact and forethought with there tokens.

I support this, both the angle and the change. "Do whatever for $200" is the ultimate shopping cart test for developers. If you've been burning tokens on performative agentic nonsense, you've actively contributed to this change.

@work_____ Broseph, Codex is a better coder. Claude is a better coworker aka long herizon terminal tasks. OpenClaw is the new benchmark and other models are coming.

Bros openclaw without Claude is shit. I can tell you bc for 2 months I’ve tried OpenAI $200 and Claude $200 subs. OpenAI is the bitch of Claude. OpenAI is good to do some cleaning and smol cooking and that’s it. Important vibe coding is for Claude and Claude only.

@andraskindler VPN

Just got a new macbook What should I install first?

@rasbt @steipete @openclaw 100% tho

@steipete @openclaw This is the true replacement of SWE-Bench Verified & Tau2 😆

Interesting benchmark on which model is best for @openclaw pinchbench.com

@vchennai2 @patmatthews @ycombinator @ArdentAI ha you say solo and the say “we let..” you and which agents.

I got into @ycombinator solo After 7 rejections Before the batch: -$2.1M raised alone -Sold 200K ARR -6M+ views I'm building the first database sandbox @ArdentAI We let you infinitely clone any Postgres DB in <6s so coding agents can test code on a 1:1 of prod Time to win