Dyslexia affects both reading and writing. It's an inherited difference in brain connection structure that can cause difficulty in associating the phonemes or sounds with letters and letter combinations. It provides advantages in three dimensional visualization and enables athletes to be ambidextrous.
Bennett offered a number of suggestions for improving Claude Code, three stood out:
A markup tool: I wish I could easily point to things: screenshots, UI glitches, layout issues, or share a whiteboard with Claude. Currently, all of these must be described completely in text. Often twice. Sometimes three times. I’ve tried uploading a marked-up screenshot, but with mixed success. A markup tool would be enormously helpful.
We need a larger context window. In regular Claude, the current 200k is being upgraded, in beta test now, at the 1M mark! But I think Claude Code’s limits are much lower. I see errors like Error: File content (27931 tokens) exceeds maximum allowed tokens (25000); 25,000 is pretty small when dealing with entire source code files.
Unprompted Git commits (This is the worst surprise.) Claude Code sometimes commits to Git without being asked. And even worse: it may commit before the change is tested; I don’t like checking in obviously broken code.
In a technical environment the first step is probably to write your ideas down. Sleep on it, review, and then share with a few people you trust for candid feedback. From there you can share more widely, fine tune and adjust, or realize that you mis-assessed.
I think in the early stages, the intent is self-rescue: prevent yourself from going off half-cocked. I see it as a way to gather feedback from trusted peers who have knowledge of the problem area. If it proves out, then they can help you sell it, but my primary intent was to encourage a "measure twice, cut once" approach instead of running with an idea without adequate preparation. If you cannot find at least three things that might go wrong with your approach, you probably have not thought it through. The intermediate step between soliciting feedback and "pre-wiring" is a pre-mortem, where you actively solicit potential failure modes and stumbling blocks.
"Regardless of technological innovations, the raw materials in question [wood, coal, oil] have never yet been obsolete. Exceptions to this rule are exceedingly rare. Whale oil offers a unique example of the disappearance of an energy source. Or I could mention sheep’s wool, whose use has declined by one-third since the 1950s, replaced by synthetic fibres. Or that of asbestos, now that it’s banned."
Q: Why has the history of energy been misrepresented like this?
J.-B. F.: The reason is simple: historians tend to look at energy from an economic viewpoint, seeking to understand the roots of industrialisation and growth. To that end, they convert those tonnes of wood, coal and petroleum into energy units, and examine the evolution of the mix in relative terms. So in the industrialised countries in 1900, the energy contribution of wood, for example, did indeed become negligible compared with that of coal.
Yet in terms of trees, biodiversity and climate, it’s absolute values that count, and the number of trees felled has never stopped increasing. Moreover, historians have not studied the interrelationships between energy sources – for example, all the wood needed to mine coal or all the coal necessary to extract and use petroleum.
Q: What solutions do you propose?
J.-B. F.: This is the key question that my book does not answer at all. What climate policy should we pursue once we realise that carbon neutrality is largely an illusion, and that we can slow down but probably not stop climate change?
There are also in-person Bootstrapper Breakfast meetings in Las Vegas, San Francisco, and Silicon Valley.
These events are for “entrepreneurs who eat problems for breakfast.”®
reply