Robert Pratt

[Robert Pratt @bponghneng]

@warpdotdev The extension failed to start. Root cause: Swift was using compat libs on the remote MacOS system. Fixed that, and now I have remote file browsing working in Warp over SSH. I probably would not have figured that out after an hour of poking around pre-agents and pre-Warp!

[Robert Pratt @bponghneng]

One of those "what did we do before agents?" moments:

I started using @warpdotdev nine months ago and two months ago promoted it to my primary software engineering tool. I had moved over from VS Code, and the biggest pain point was SSH remote file browser support. Now fixed!

[Robert Pratt @bponghneng]

@warpdotdev After updating Warp, I tried installing the new SSH feature, which uses a downloaded extension on the remote SSH system. The install failed, so I threw the error at GPT-5.5 with high reasoning. It tried a few paths, then manually downloaded and installed the extension.

[Robert Pratt @bponghneng]

@heyhve_ Totally. This one is built for human signoff before fix. My experience is that something like a Ralph loop until there are zero issues or only nitpicks is highly susceptible to drift.

[Robert Pratt @bponghneng]

I published my take on code review automation after several months of using review tools and building my own.

It has two workflows:
- review: identify PR/MR issues
- fix: apply selected fixes with human judgment in the loop

https://t.co/2b0BBXO2CV

[Robert Pratt @bponghneng]

The human step matters.

Some agent-flagged issues should not be fixed. Some low-priority or low-confidence issues are still worth fixing in context.

The goal is not autonomous review. It’s more thorough, consistent review with a human keeping the agent on track.

[Robert Pratt @bponghneng]

Building a custom n8n agent that takes any article → scrapes it → sorts into 6 AI categories → generates category-specific summaries (e.g., for coding tools, I want author expertise assessment included).

Getting the classification prompts right is trickier than it looks.

[Andrew Ng @AndrewYNg]

On Monday, a United States District Court ruled that training LLMs on copyrighted books constitutes fair use. A number of authors had filed suit against Anthropic for training its models on their books without permission. Just as we allow people to read books and learn from them to become better writers, but not to regurgitate copyrighted text verbatim, the judge concluded that it is fair use for AI models to do so as well.

Indeed, Judge Alsup wrote that the authors’ lawsuit is “no different than it would be if they complained that training schoolchildren to write well would result in an explosion of competing works.” While it remains to be seen whether the decision will be appealed, this ruling is reasonable and will be good for AI progress. (Usual caveat: I am not a lawyer and am not giving legal advice.)

AI has massive momentum, but a few things could put progress at risk:
- Regulatory capture that stifles innovation, including especially open source
- Loss of access to cutting-edge semiconductor chips (the most likely cause would be war breaking out in Taiwan)
- Regulations that severely impede access to data for training AI systems

Access to high-quality data is important. Even though the mass media tends to talk about the importance of building large data centers and scaling up models, when I speak with friends at companies that train foundation models, many describe a very large amount of their daily challenges as data preparation. Specifically, a significant fraction of their day-to-day work follows the usual Data Centric AI practices of identifying high-quality data (books are one important source), cleaning data (the ruling describes Anthropic taking steps like removing book pages' headers, footers, and page numbers), carrying out error analyses to figure out what types of data to acquire more of, and inventing new ways to generate synthetic data.

I am glad that a major risk to data access just decreased. Appropriately, the ruling further said that Anthropic’s conversion of books from paper format to digital — a step that’s needed to enable training — also was fair use. However, in a loss for Anthropic, the judge indicated that, while training on data that was acquired legitimately is fine, using pirated materials (such as texts downloaded from pirate websites) is not fair use. Thus, Anthropic still may be liable on this point. Other LLM providers, too, will now likely have to revisit their practices if they use datasets that may contain pirated works.

Overall, the ruling is positive for AI progress. Perhaps the biggest benefit is that it reduces ambiguity with respect to AI training and copyright and (if it stands up to appeals) makes the roadmap for compliance clearer. This decision indicates it is okay to train on legitimately acquired data to build models that generate transformational outputs, and to convert printed books to digital format for this purpose. However, downloading from pirate sites (as well as permanently building a “general purpose” library of texts, stored indefinitely for purposes to be determined, without permission from the relevant copyright holders) are not considered fair use.

I am very sympathetic with the many writers who are worried about their livelihoods being affected by AI. I don‘t know the right solution for that. Society is better off with free access to more data; but if a subset of people is significantly negatively affected, I hope we can figure out an arrangement that compensates them fairly.

[Original text: https://t.co/kxcCgL4tpH ]

[Vaughn Vernon @VaughnVernon]

***MORE***
My Experiences with Human Software Developers vs. AI.

When you describe the socio-technical relationship between business and software people, don't forget the part where the *typical* software developers tell the business people that they're wrong. Only then do they interpret as they prefer and transform that into inscrutable artifacts that are hard to change. When it's finally understood that the developers were wrong, the best possible outcome for developers is for the business people not to understand what happens inside.

Besides several decades of working for business people, I've also been the business person specifying, facilitating, and paying real money for it. The majority of experiences have been right up/down there with the worst in my career.

Along with being told that I'm wrong, that I don't know what I'm talking about, and being completely ignored, the artifacts are inscrutable. I'm not a slouch developer, but I don't want to touch the code that I've paid to have written for me.

I've described my new relationship with AI as unpredictable to the point of inducing stress (e.g., I've lost count of how many times it's deleted all hundreds of source code files, replacing them with a few files), the AI isn't 1/1000th as bratty as human software developers. (I'm polite, and it apologizes.) And I can read its code when I've given it clear specifications 💰

I want to be clear. My results are not from vibe coding. I am meticulously specifying both architecture and modeling the domain. The difference is that I'm writing far less code myself, but the outcome is pleasing.

Business people will not vibe code anywhere near the same excellent outcomes. Neither will average software developers. The AI will win out and demoralize them.

[I suggest reading Tudor's full post on LinkedIn.]

[Robert Pratt @bponghneng]

Hmm, I think I've spent time in all of the quadrants.

[Tim Ottinger @tottinge]

I queued up a tweet on this, but the image is too darned fine to not post it here now:

[I Am Devloper @iamdevloper]

teacher: if I have 0.1 apples in one hand and 0.2 apples in the other hand, how many apples do I have?
javascript: 0.30000000000000004

[Josh Pollock 🌱🌵🌲 @Josh412]

The CDC no longer requires developers to fully sanitize user input before performing a SQL query.

[Robert Pratt @bponghneng]

@getolympus 22nd in the series. Last of the earlier images to get a rework. #dallas #architecturephotography #abstractphotography #urbanphotography

[Robert Pratt @bponghneng]

@getolympus 39th in the series. Geometrics found among a tower edge of Fountain Place in the city center of Dallas, TX. #dallas #architecturephotography #abstractphotography #urbanphotography

[Robert Pratt @bponghneng]

@getolympus 7th in the series. Final redux of one of the early images. If you haven't noticed just yet, I often think of Josef Albers when I work on these images. #dallas #architecturephotography #abstractphotography #urbanphotography

[Robert Pratt @bponghneng]

@getolympus
40th in the series. Grids of electric-color foreground lines recur in my imagination. Credit seeing "Tron" at an early age. #dallas #architecturephotography #abstractphotography #urbanphotography

[Extinguished Engineer @ExtinguishedEng]

Hexagonal architecture makes less sense if we see our entire application in terms of CRUD, where the application gets some entity, modifies it, and then says, "Here's my updated version of that entity." 1/

[Robert Pratt @bponghneng]

@getolympus 21st in the series. Another redo of a one of the early images in the series. This one is Energy Plaza in the city center of Dallas, TX. #dallas #architecturephotography #abstractphotography #urbanphotography

[Robert Pratt @bponghneng]

@getolympus 42nd in the series. An abstract view of geometrics on Fountain Place in the city center of Dallas, TX. #dallas #architecturephotography #abstractphotography #urbanphotography

[Robert Pratt @bponghneng]

@getolympus 37th in the series. Minimal view of a façade of Fountain Place in the city center of Dallas, TX. #dallas #architecturephotography #abstractphotography #urbanphotography