Thomas Cothran

Clojure Engineer

[Thomas Cothran @thomas_cothran]

A #Clojure story about the REPL in production.

We had the opportunity at work to support NFL broadcast. We take in game data, run millions of simulations, and project outcomes of plays/games.

Timeline: 2 weeks. 😅
Margin for error: live broadcast, real-time, so very tight

[Thomas Cothran @thomas_cothran]

@iruletheworldmo There is the same hysteria any time information becomes more widely available. The printing press, the internet, open source encryption.

The apocalyptic predictions don't come true. But they do serve entrenched, powerful interests.

[Thomas Cothran @thomas_cothran]

@shashj https://t.co/fSXzLZ4quk

[Thomas Cothran @thomas_cothran]

@shashj It's almost as if there's some third party involved, one who may have been aware of Anthropic's scaremongering, have lacked the competence to accurately assess the risks, and had the power to do something about it.

[Thomas Cothran @thomas_cothran]

@shashj So guard-railed you can't get help with your biology homework.

Meanwhile, other LLMs with more or less equivalent capabilities are available at much lower prices.

Anthropic fear-mongered about Mythos to create hype about its capabilities and cover for their lack of compute

[Thomas Cothran @thomas_cothran]

@unclebobmartin This aged well

[Thomas Cothran @thomas_cothran]

@iruletheworldmo This is a parody account

[Arnaud Bertrand @RnaudBertrand]

By the way, public service announcement: if you're one of the numerous people posting about Anthropic's dystopian ways and you're thinking about getting Claude to help you write that post... don't!

Another one of their terms is that you may not use Claude to do anything that "exposes [Anthropic to] reputational harms" 👇

And, if you do, under the - extremely unusual - clause 13 of their terms (https://t.co/z43rJNkvZu), you have PRE-AGREED, by using Anthropic (and accepted their terms), that the harm you've done is irreparable, that you won't oppose Anthropic injunction, and they don't need to prove actual damage.

They can simply go to a judge in a friendly jurisdiction (and of course, their terms precise that any dispute "will be resolved exclusively in the state or federal courts located in San Francisco, California") and:
a) file an injunction that shuts you down
b) make you pay for everything since under section 11 of their terms you agree to indemnify Anthropic for "any and all liabilities, claims, damages, expenses (including reasonable attorneys' fees and costs), and other losses arising out of or related to your breach or alleged breach of these Terms."

In other words, if you use Claude to help you talk shit about Anthropic publicly, their terms say you pay their lawyers to go after you and you've already pre-agreed you've lost the case.

Oh, and cherry on the cake: in the odd case the judge were like "are you crazy, this is insanely abusive, you Anthropic are the ones at fault here," according to their terms Anthropic's maximum liability is... $100.

[Thomas Cothran @thomas_cothran]

@unclebobmartin By far, most money in the money supply (M2) is created by private banks. This is basic to economic literacy.

https://t.co/t7rzHTXJzF

[Dan Lentz @danlentz]

clj-xref 0.1.0 — LLM-friendly cross-reference database for Clojure code

https://t.co/8EEBXJvJjR

Instead of feeding your entire source tree to an AI coding assistant, clj-xref lets it query the dependency neighborhood of the function it's working on — just the relevant code, in far fewer tokens. Built on clj-kondo static analysis, it generates a queryable xref database with **lein xref ** or clj -T:xref generate, then answers questions from the REPL:

(xref/who-calls db 'myapp.orders/process-payment)
(xref/calls-who db 'myapp.web/handler)
(xref/who-implements db 'myapp.protocols/Billable)
(xref/unused-vars db)
(xref/call-graph db 'myapp.core/main {:depth 3})

Returns plain maps and vectors — compose with filter, map, group-by, or anything else.

How much context does it save? Well, that is still an experiment, but Measured on clj-xref's own source:

│ Target │ Whole-tree │ Xref-guided │ Reduction │

│ core/index │ 8 files, 32K chars │ 1 file, 6K chars │ 80% │

│ analyze/transform-analysis │ 8 files, 32K chars │ 1 file, 7K chars │ 78% │

│ leiningen.xref/xref │ 8 files, 32K chars │ 3 files, 11K chars │ 66% │

│ analyze/classify-kind │ 8 files, 32K chars │ 1 file, 7K chars │ 78% │

Also useful for: dead code detection, codebase visualization (DOT/Graphviz output), impact analysis in CI, and REPL-driven exploration of unfamiliar code.

Query API: who-calls, calls-who, who-references, who-macroexpands, who-implements, who-dispatches, unused-vars, call-graph, apropos, ns-deps, ns-dependents. Inspired by SBCL's sb-introspect and Smalltalk's senders/implementors.

Built entirely on clj-kondo — please consider sponsoring @borkdude: https://t.co/qI5yDu4mWn

[Thomas Cothran @thomas_cothran]

One of the most useful features of LLMs - producing minimal repro cases for tricky bugs. Takes something that could be an hour+ down to a few minutes

[Thomas Cothran @thomas_cothran]

@unclebobmartin You went from confidently declaring what the legal landscape was on ICE/4A to declaring you don't know what "probable cause" or "oath or affirmation" means.

Next you may want to check out who the Constitution gives the power to declare war. The answer might surprise you.

[Thomas Cothran @thomas_cothran]

@unclebobmartin What interpretation are you giving to the words probable cause or oath or affirmation exactly?

I know not everyone is a lawyer, but declaring something any 2L would know is wrong so confidently that uninformed readers might be led to believe you have some knowledge ... not great

[Thomas Cothran @thomas_cothran]

@unclebobmartin Were you informed by the text of the 4A you would see the words "warrant", "probable cause", "oath or affirmation"

Were you then to look at an I-200, you would see neither an oath or affirmation, nor any probable cause requirement.

You are confidently and incorrigibly incorrect

[Thomas Cothran @thomas_cothran]

@unclebobmartin The 4A requires probable cause, an oath or affirmation, and a warrant signed by a judicial officer. *None* of these are the case with a LEO signing I-200 or I-205 forms.

I learned this in my conlaw class years ago. You're informed by ... what exactly?

[Thomas Cothran @thomas_cothran]

@unclebobmartin Nobody tell him who's in charge of Iran now

[Thomas Cothran @thomas_cothran]

@jpodhoretz @PhillipsPOBrien For @jpodhoretz the war ought to be won by bombing more elementary schools.

Unsurprising for someone who defends the wisdom Iraq war and said it would have been won if only we had more mass killing.

[Thomas Cothran @thomas_cothran]

@sama The executive branch believes that is legal to kill civilians suspected of drug offenses, and has been doing so for months.

Does OpenAI's agreement prohibit using ChatGPT in strikes on civilians where the exec branch argues it is lawful?

[Thomas Cothran @thomas_cothran]

@lydiahallie official opencode support

[Thomas Cothran @thomas_cothran]

@trq212 @burkov Could be Claude Code regressions.

Codex subscriptions work in Opencode now (officially).

I made the switch and the results are better than Claude Code.