if you could look clearly at the progress from 2020 to 2023, as someone like Gwern did, and from 2023 to 2024 with the invention of reasoning modes, then it was not that hard to understand what would happen in late 2025. Opus 4.5 was not a surprise to anyone who was actually paying attention.

But people (including the author) still mistake the current state as a stable state and future gains as incremental. he says

“I am not asserting that all code will eventually be AI-generated to spec, bypassing human understanding”

I AM asserting that, and it’s incredibly easy to do so.

The question of “when” is separate.

What was especially useful (and unusual in IPC systems it seems) was the property that message channels outlive all readers and writers, and even survive reboots, because they're just files. For local IPC you don't need a broker or server process.

All the engineers who ever worked at Oblong loved Plasma, so I've recreated and updated it, as Plasmite.

It's written in Rust and the message format is JSON, but it's fast because it's based on lite3 (https://lite3.io/index.html), a really cool project you should also check out.

Bindings for Python, Go, Node, and C, but you can also get a lot done with just the CLI tools. The basic commands are - "feed" (to write) - "follow" (to tail) - "fetch" (to read one) - "duplex" (to have a 2-way session)

I think duplex could be great for agent-agent communication, but I haven't tried this much yet. If you do, let me know!

Comments URL: https://news.ycombinator.com/item?id=47511435

Points: 9

# Comments: 1

Backlogs are better off being represented as a task dependency graph, rather than a heap of markdown files, because a nice graph of tasks:

- helps agents focus - gives better observability of partial progress - supports parallelization better - works well with subagents

Also, the act of planning a backlog *makes the design better* because in the act of planning, you learn things about your own design. This is as true for agents as it is for humans.

The planning modes of claude and codex blur the distinction between specs and backlog and land you in a heap of markdown files.

Steve Yegge diagnosed all this, and his solution was the beads CLI -- a brilliant idea. You just tell your agent to use this CLI for planning instead of their own plan mode (all you do is add a one-liner in your AGENTS.md file), and you get an instant planning upgrade.

However, I found the implementation of beads a little bit hairy and flaky, when I tried to use it in production. And the performance made me sad.

So I've reimplemented the core concepts, as ergo.

It's very fast (5-15x faster than beads), very robust, and rigorously simple. There is no daemon running, no SQL database, it's self-healing, and the plans are stored as JSONL, so if all else fails you can just reconstruct your plans from scratch.

Implementation:

- Plans live in `.ergo/` in your repo root (or your dir of choice). JSONL was chosen because it's git-friendly & easy to resolve merge conflicts around. - ergo is concurrency-ready if you're into agent swarms. Concurrent writes are serialized with flock(2). Multiple agents can race to claim tasks — exactly one wins, others fail fast. - Do operations in huge plans of 1000+ tasks in ~15ms on a Macbook Air. - The help text doubles as the agent manual. I spent a lot of time on `ergo --help` and `ergo quickstart` because they're the primary interface for agents.

I've been using ergo heavily for a couple months and I believe it's solid enough for anyone to use, would love your feedback.

Comments URL: https://news.ycombinator.com/item?id=47010903

Points: 2

# Comments: 0

If we were to treat Apple and Meta equivalently, we would have spent a decade tallying Apple’s R&D investments as “losses”.

Does this imply that the blockchain ledger might have greater utility in areas where there is weak or no state control? Basically, the criminal underworld and black markets. What's been written about this?

Magnopus has deep expertise in AR, VR, mixed reality, and virtual production. We’re passionate about creating experiences that bridge the physical and digital worlds in ways people haven’t experienced yet.

We're looking for more than one front-end web developer to join our new product team, working out of the brand new Magnopus HQ in downtown LA (open July-ish). https://apply.workable.com/magnopus-2/j/183A32ACD7/

- You have multiple years of experience shipping production-quality React applications that interface with back-end services -- link us to your portfolio

- You’re strong on the fundamentals of Javascript (preferably also Typescript), React (and React Hooks), CSS

- You are familiar with modern frameworks and can demonstrate your ability to choose wisely among them

- B.S. degree in Computer Science preferred but not required

- 5+ years full time work experience

Magnopus is an experience company with deep expertise in AR, VR, mixed reality, and virtual production. We’re passionate about creating experiences that bridge the physical and digital worlds in ways people haven’t experienced yet.

Come help us build a new spatial product & platform touching on numerous tech stacks and all major platforms.

- Full stack web developer: https://apply.workable.com/magnopus-2/j/183A32ACD7/

- Spatial UI Developer: https://apply.workable.com/magnopus-2/j/2F5DB975A6/

The crew at Magnopus are smart, they are good communicators, and they are gracious. So ideally you're like that, too. These are growth roles in a growing company in a growing team; you will not only ensure we get good results up front, but you will also have an opportunity to set best practices that will serve us for years to come.

(Other positions are also open, including Security Engineer.)

Magnopus is an experience company with deep expertise in AR, VR, mixed reality, and virtual production. The company was founded by a team of Oscar-winning artists, designers, and engineers from games, films, and other unusual places. We’re passionate about creating rich experiences that bridge the physical and digital worlds in ways people haven’t experienced yet.

We are building a new product and platform encompassing front end, back end, and middleware -- and covering desktop, web, mobile, VR, and AR.

- Full stack web developer: https://apply.workable.com/magnopus-2/j/183A32ACD7/

- SDET (Software Development Engineer in Test): https://apply.workable.com/magnopus-2/j/DFCEF294C0/

- Spatial UI Developer: https://apply.workable.com/magnopus-2/j/2F5DB975A6/

The crew at Magnopus are smart, they are good communicators, and they are gracious. So ideally you're like that, too. These are all growth roles in a growing company in a growing team; you will not only ensure we get good results up front, but you will also have an opportunity to set best practices that will serve us for years to come.

(Other positions are also open, including Security Engineer.)

Magnopus is an experience company with deep expertise in AR, VR, mixed reality, and virtual production. The company was founded by a team of Oscar-winning artists, designers, and engineers from games, films, and other unusual places. We’re passionate about creating rich experiences that bridge the physical and digital worlds in ways people haven’t experienced yet.

We are building a new platform encompassing front end, back end, and middleware -- and covering desktop, mobile, VR, and AR. To ensure robustness and quality in everything we build, we need a skilled Software Development Engineer in Test (SDET) to design, create, and own automated testing and test infrastructure.

https://apply.workable.com/magnopus-2/j/DFCEF294C0/

The crew at Magnopus are smart, they are good communicators, and they are gracious. So ideally you're like that, too. This is a growth role in a growing company in a growing team; you will not only ensure we get good results up front, but you will also have an opportunity to set best practices that will serve us for years to come.

(Other positions are also open, including Security Engineer.)

Oblong is a hardware and software company that has always worked at the forefront of human-computer interaction. We are looking for talented folks to help us radically evolve our next-generation meeting tool, Mezzanine.

Currently open positions:

- full-stack engineer (React, Redux, node...)

- back-end engineer (Go, GCP...)

- C++ engineer

- Security engineer

- Director of Engineering

- and more

We have a small, agile engineering team where you can move fast and make an outsized impact.

Oblong is a humane place to work. For example, the company pays 100% of health insurance costs.

email brandon at oblong

https://www.oblong.com/company/careers

Oblong is a hardware and software company that has always worked at the forefront of human-computer interaction. We are looking for talented folks to help us evolve our next-generation meeting room product, Mezzanine.

Currently open positions: full-stack engineer (React; Redux; Electron a plus) or back-end engineer (Go, etc.) We have a small, agile engineering team where you can move fast and make an outsized impact.

Oblong is a humane place to work. For example, the company pays 100% of health insurance costs.

email brandon at oblong

Oblong is a hardware and software company that has always worked at the forefront of human-computer interaction. We build meeting room products and large-scale immersive systems.

We are hiring in LA for positions in full-stack web (React, Redux, Electron a plus) as well as back-end (Go, etc.) We have a small, agile engineering team where you can make an outsized impact.

More positions will open up; we are also interested in graphics experience and experience with video or videoconferencing. Know Chromium? Graphics drivers? C++? Talk to us.

Oblong is a humane place to work. For example, the company pays 100% of health insurance costs.

email brandon at oblong

These types of workplaces are not a newly-hatched resource squeeze by our ruthless capitalist overlords. They are a physical work pattern which has been reinvented and retooled many times, and isn't going away. In most of history, "office"==="big room".

Fixed-route 19th-20th century public transportation networks are inefficient because most of the vehicles are pretty empty most of the time. This makes these systems expensive on a passenger-per-seat-mile basis. Not to mention the colossal waste of people's time spent trying to accommodate their journeys to those fixed routes (and schedules).

Example: LA's Metro system is subsidized to the tune of about 50 cents per passenger seat/mile, or was as of 2009 (http://www.newgeography.com/content/002361-los-angeles-metro...)

In 10-15 years, transport companies like Lyft/Uber/Didi will be able to offer municipalities a much better deal than that. An electric car or van without a driver in it, with an operating lifetime of a million miles, will cost on the order of 20-40 cents a mile to run. Get 3-8 paying customers into the vehicle, and there is absolutely no need for a 50 cent per passenger seat mile subsidy.

1. extremely well defined problem domains

2. incrementalism

For lots of engineers, those 2 points describe their job. Lots of (incremental) value is created that way.

But I'd argue that the larger gains accrue to those companies -- or those cities -- where work is being done in non-well-defined problem domains. And not incrementally, but by big leaps, experiments, and messy failures.

In fact, you could go farther and argue that in a competitive landscape, companies will quickly match each other in their performance in well-defined problem spaces, and the only way to actually compete at the margins is in problem areas that are, by definition, newer and not well defined.

Helmet paranoia DOES reduce the odds of exercise. On a group level, at the level of public health, that's PRECISELY what it does. I encourage you to read about it, that's why they call it the helmet paradox. This piece is a nice summary.

http://www.howiechong.com/journal/2014/2/bike-helmets

Bike helmets protect our heads, but they also do the following:

1. increase your chances of getting in an accident

2. discourage cycling

Here's what to do:

1. If you want to be healthy, ride a bike instead of driving.

2. If you want to be even healthier, consider maybe putting on a helmet, but be mindful it doesn't increase your risk-taking behavior.

I choose option 2, but option 1 is also great.

http://ktla.com/2013/09/23/gov-brown-signs-law-requiring-car...

But it doesn't change the truth of the sentence: "The exercise benefits of biking do way more to increase your life expectancy than skipping a helmet does to reduce it."

The person who is truly taking a risk is the person who doesn't ride a bike because of the perception that it's dangerous, and continues to be sedentary and drive a car to work. This person is literally reducing their life expectancy by years. (See links I posted elsewhere in the thread.)

You need to be able to think about conditional probability.