Dan Tocchini IV

If I be more noble than you, it is b/c I have suffered more. intersect(Jesus, code, design) Berkeley EE Physics @thegrid

[Dan Tocchini IV @d4tocchini]

1 year ago today, my son Aeon died when a 120ft redwood snapped 40ft in the air & free fell on our house crushing him just 5ft away from me.

From the abyss of suffering, I return new born, having shed my skin. Aeon's death saved my life.

This is my testimony.

An evergreen 🧵

[Dan Tocchini IV @d4tocchini]

@QuixiAI Good to know

[Dan Tocchini IV @d4tocchini]

@gietzschean Katechon vibes

[Dan Tocchini IV @d4tocchini]

@bunjavascript Seems so obvious now, the types are already there, fuck yes they should be real.

[Hari @HarivanshRathi]

on ai psychosis:

as i watch engineers fall deeper into the belief that an amalgamation of mathematical probabilities somehow understands their codebase better than they do, i find myself thinking back to a time when software wasn’t built for hypergrowth, but simply to do x without inventing y.

ai seems almost fundamentally opposed to this philosophy. ask it to do x and it will eagerly invent y and z before it has even tried to understand x.

the danger isn’t that ai writes bad code- it’s that it makes writing unnecessary code 100% free

and the human condition is such that some of us will always prefer the fast, steep gains of ai, even when it does a bajillion unrelated things to accomplish something that could have been done without changing anything else.

so, somewhat paradoxically, the quality of software may keep declining for as long as ai keeps getting better. the cheaper complexity becomes to create, the less incentive there is to understand or avoid it.

try to preserve this craft created by our ancestors
write a LOC by hand today

[Dan Tocchini IV @d4tocchini]

@greghunkins @pmarca @MageSpace_ So good

[Dan Tocchini IV @d4tocchini]

@QuixiAI @raysan5 Omg, have you played Geometry Wars?! I love that game

[Dan Tocchini IV @d4tocchini]

@sophiadew Yes

[Dan Tocchini IV @d4tocchini]

@xenocosmography "He said to him, ‘If they do not hear Moses and the Prophets, neither will they be convinced if someone should rise from the dead.’"

[Dan Tocchini IV @d4tocchini]

@Nick_Davidov Anthropic cockblocks API usage with blunders like this.

I love Fable, but at a certain point interoperability trumps slightly higher benchmark scores.

https://t.co/7WUbz7sm6J

[Armin Ronacher ⇌ @mitsuhiko]

We're currently still evaluating how to deal with Fable 5.1's inference restrictions. Anthropic is restricting mid-conversation model switches (you lose all reasoning) and we still need a day or two to figure this out. For now it only applies to new organizations. https://t.co/meQwcfTXJ7

[Dan Tocchini IV @d4tocchini]

@rikarends Pascal's wager for accelerationism

[Dan Tocchini IV @d4tocchini]

@Vtrivedy10 "Nothing is now unconscious if your data search commands are powerful enough.”

– Terence McKenna in Riding Range w/ Marshall McLuhan

https://t.co/yMMUYR4IQY

[Viv @Vtrivedy10]

from Astra blog - Codex compaction is already great, never think about it, have multi-week threads

turns out avoiding destructive (unrecoverable) summarization actions + letting a very smart model search over previous context…just works

scaling + improving general search works, bitter lesson is inescapable

[Dan Tocchini IV @d4tocchini]

@thdxr Read this & thought, ooh, whats the golang drama

https://t.co/P1ENckTWjh

[dax @thdxr]

working on OpenCode Go has been the hardest thing i've ever worked on

- most competitive space on the planet
- massive levels of fraud
- trying to achieve impossible economics

there's a 3% chance we make it all work

and if we do everyone will talk about it like it was always inevitable

and only a handful of people will ever understand how it very much was not

[Dan Tocchini IV @d4tocchini]

@rikarends Yes, I have been playing around with it! Very cool

That's also a testament to the foundation you laid

[Dan Tocchini IV @d4tocchini]

@rikarends you undersell the value of knowing what magic spells are appropriate to utter and when!

[Dan Tocchini IV @d4tocchini]

@rikarends yea, its addicting

[Dan Tocchini IV @d4tocchini]

@JakedChinatown Nice

Even lifted the intro song from Hypernormalisation

If AI gives us more Adam Curtis, I am happy

https://t.co/5Viq4TlwN4

[Dan Tocchini IV @d4tocchini]

There is no alpha in training your own foundation model.

The best techniques will get discovered, copied and eventually incorporated into open-weight models like GLM.

There is no alpha in building your own vertical app.

The best features will get discovered, copied and eventually incorporated into countless vibe coded apps like bespoke video editors & CRMs.

Yet, unicorns abound across harness, model & SaaS cos. Alpha is a function of execution & timing, not of product category.

https://t.co/WGcEOx4zkT

[Sherwood @shcallaway]

There is no alpha in building your own agent harness.

The best techniques will get discovered, copied, and eventually incorporated into open-source harnesses like Pi.

A few months after that happens, these techniques will be made obsolete entirely as labs release smarter models + model providers integrate upwards (pulling more of the harness behind the model API endpoint).

Am I wrong? Please let me know 🧐

[Dan Tocchini IV @d4tocchini]

@Joshua_Ariza Intellectualism is a common cover-up for fear of direct experience.
— Carl Jung

[Dan Tocchini IV @d4tocchini]

@JakedChinatown Nice

Even lifted the intro song from Hypernormalisation

If AI gives us more Adam Curtis, I am happy

https://t.co/5Viq4TlwN4