Hacker News: philipkglass

New comment by philipkglass in "The noise we make is hurting animals. Can we learn to shut up?"

philipkglass — Thu, 16 Apr 2026 21:16:41 +0000

This was interesting enough to prompt an investigation. It looks like the idea comes from this letter that he wrote late in life:

https://philosophics.blog/2021/08/01/the-futility-of-words/

I don't know German so I can't search for the original with any facility.

New comment by philipkglass in "CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous"

philipkglass — Wed, 15 Apr 2026 18:37:38 +0000

I think that we're getting there. I put together a workstation in early 2023 with a single 4090 GPU. I did it to run things like BERT and YOLO image classifiers. At that point the only "open weights" LLM was the original Llama from Meta, and even that was open-weights only because it was leaked. It was a very weak model by today's standards.

With the same hardware I now get genuine utility out of models like Qwen 3.5 for categorizing and extracting unstructured data sources. I don't use small local models for coding since frontier models are so much stronger, but if I had to go back to small models for coding too they would be more useful than anything commercially available as recently as 4 years ago.

New comment by philipkglass in "The Future of Everything Is Lies, I Guess: Safety"

philipkglass — Mon, 13 Apr 2026 17:56:21 +0000

In short, the ML industry is creating the conditions under which anyone with sufficient funds can train an unaligned model. Rather than raise the bar against malicious AI, ML companies have lowered it.

This is true, and I believe that the "sufficient funds" threshold will keep dropping too. It's a relief more than a concern, because I don't trust that big models from American or Chinese labs will always be aligned with what I need. There are probably a lot of people in the world whose interests are not especially aligned with the interests of the current AI research leaders.

"Don't turn the visible universe into paperclips" is a practically universal "good alignment" but the models we have can't do that anyhow. The actual refusal-guards that frontier models come with are a lot more culturally/historically contingent and less universal. Lumping them all under "safety" presupposes the outcome of a debate that has been philosophically unresolved forever. If we get hundreds of strong models from different groups all over the world, I think that it will improve the net utility of AI and disarm the possibility of one lab or a small cartel using it to control the rest of us.

New comment by philipkglass in "Seven countries now generate nearly all their electricity from renewables (2024)"

philipkglass — Sun, 12 Apr 2026 14:51:33 +0000

At present, those tandem cells are still experimental. Nobody is manufacturing them on gigawatt scale like for other solar cell technologies.

First criticality for Indian fast breeder reactor

philipkglass — Tue, 07 Apr 2026 18:16:45 +0000

Article URL: https://www.world-nuclear-news.org/articles/first-criticality-for-indian-fast-breeder-reactor

Comments URL: https://news.ycombinator.com/item?id=47679228

Points: 3

# Comments: 0

New comment by philipkglass in "Show HN: I Built Paul Graham's Intellectual Captcha Idea"

philipkglass — Mon, 06 Apr 2026 16:53:28 +0000

At least one of the test questions was just a screen shot from a tweet. It was difficult to read. I'd suggest extracting text from screen shots with OCR. Apple has built-in functionality for this on their operating systems with Live Text. There are strong open source systems based on small vision language models for this, too. The one I have been recommending lately is GLM-OCR:

https://github.com/zai-org/GLM-OCR

It's fast and can run even on low-resource computers.

---

Does this CAPTCHA actually resist computers? I didn't try feeding the questions I got to an LLM, but my sense is that current frontier models could probably pass all of these too. Making generated text pass the pangram test is simple enough for someone actually writing a bot to spin up automated accounts.

New comment by philipkglass in "Solar and batteries can power the world"

philipkglass — Fri, 03 Apr 2026 17:47:13 +0000

Unlike the US, Vietnam is a net importer of fuel. It imports over 40 million tons of coal per year:

https://statbase.org/data/vnm-coal-imports/

It also started importing liquid natural gas in 2023.

But it has abundant sunlight, access to low cost Chinese solar panels that will produce electricity for decades instead of being burned once, and a substantial domestic photovoltaic manufacturing industry of its own.

"Renewable Energy Investments in Vietnam in 2024 – Asia’s Next Clean Energy Powerhouse" (June 2024)

https://energytracker.asia/renewable-energy-investments-in-v...

In 2014, the share of renewable energy in Vietnam was just 0.32%. In 2015, only 4 megawatts (MW) of installed solar capacity for power generation was available. However, within five years, investment in solar energy, for example, soared.

As of 2020, Vietnam had over 7.4 gigawatts (GW) of rooftop solar power connected to the national grid. These renewable energy numbers surpassed all expectations. It marked a 25-fold increase in installed capacity compared to 2019’s figures.

In 2021, the data showed that Vietnam now has 16.5 GW of solar power. This was accompanied by its green energy counterpart wind at 11.8 GW. A further 6.6 GW is expected in late 2021 or 2022. Ambitiously, the government plans to further bolster this by adding 12 GW of onshore and offshore wind by 2025.

These growth rates are actually much faster than growth rates in the US.

New comment by philipkglass in "Google releases Gemma 4 open models"

philipkglass — Fri, 03 Apr 2026 14:54:33 +0000

When you self-host, there's still a client/server relationship between your self-hosted inference server and the client that manages the processing of individual pages. You can get timeouts depending on the configured timeouts, the speed of your inference server, and the complexity of the pages you're processing. But you can let the client retry and/or raise the initial timeout limit if you keep running into timeouts.

That said, this is already a small and fast model when hosted via MLX on macOS. If you run the inference server with a recent NVidia GPU and vLLM on Linux it should be significantly faster. The big advantage with vLLM for OCR models is its continuous batching capability. Using other OCR models that I couldn't self-host on macOS, like DeepSeek 2 OCR or Chandra 2, vLLM gave dramatic throughput improvements on big documents via continuous batching if I process 8-10 pages at a time. This is with a single 4090 GPU.

New comment by philipkglass in "Google releases Gemma 4 open models"

philipkglass — Thu, 02 Apr 2026 22:04:54 +0000

I also get connection timeouts on larger documents, but it automatically retries and completes. All the pages are processed when I'm done. However, I'm using the Python client SDK for larger documents rather than the basic glmocr command line tool. I'm not sure if that makes a difference.

New comment by philipkglass in "Google releases Gemma 4 open models"

philipkglass — Thu, 02 Apr 2026 20:47:24 +0000

If you have a basic ARM MacBook, GLM-OCR is the best single model I have found for OCR with good table extraction/formatting. It's a compact 0.9b parameter model, so it'll run on systems with only 8 GB of RAM.

https://github.com/zai-org/GLM-OCR

Use mlx-vlm for inference:

https://github.com/zai-org/GLM-OCR/blob/main/examples/mlx-de...

Then you can run a single command to process your PDF:

  glmocr parse example.pdf

  Loading images: example.pdf
  Found 1 file(s)
  Starting Pipeline...
  Pipeline started!
  GLM-OCR initialized in self-hosted mode
  Using Pipeline (enable_layout=true)...

  === Parsing: example.pdf (1/1) ===

My test document contains scanned pages from a law textbook. It's two columns of text with a lot of footnotes. It took 60 seconds to process 5 pages on a MBP with M4 Max chip.

After it's done, you'll have a directory output/example/ that contains .md and .json files. The .md file will contain a markdown rendition of the complete document. The .json file will contain individual labeled regions from the document along with their transcriptions. If you get all the JSON objects with

  "label": "table"

from the JSON file, you can get an HTML-formatted table from each "content" section of these objects.

It might still be inaccurate -- I don't know how challenging your original tables are -- but it shouldn't be terribly slow. The tables it produced for me were good.

I have also built more complex work flows that use a mixture of OCR-specialized models and general purpose VLM models like Qwen 3.5, along with software to coordinate and reconcile operations, but GLM-OCR by itself is the best first thing to try locally.

New comment by philipkglass in "Renewables reached nearly 50% of global electricity capacity last year"

philipkglass — Thu, 02 Apr 2026 17:44:52 +0000

It takes time for statistical agencies to compile reports. I haven't yet found reports covering the growth in renewable generation (actual terawatt hours) for all of 2025. But this covers 3 quarters of the year:

https://ember-energy.org/latest-updates/solar-and-wind-growt...

In the first three quarters of 2025, solar generation rose by 498 TWh (+31%) and already surpassed the total solar output in all of 2024. Wind generation grew by 137 TWh (+7.6%). Together, they added 635 TWh, outpacing the rise in global electricity demand of 603 TWh (+2.7%).

New comment by philipkglass in "Google releases Gemma 4 open models"

philipkglass — Thu, 02 Apr 2026 17:25:23 +0000

Do you have plans to do a follow-up model release with quantization aware training as was done for Gemma 3?

https://developers.googleblog.com/en/gemma-3-quantized-aware...

Having 4 bit QAT versions of the larger models would be great for people who only have 16 or 24 GB of VRAM.

New comment by philipkglass in "Slovenia becomes first EU country to introduce fuel rationing"

philipkglass — Fri, 27 Mar 2026 21:56:16 +0000

"Lithium mining commences in Finland"

https://www.electrive.com/2026/02/12/lithium-mining-commence...

This week, the first spodumene vein was blasted from the rock at the open-pit mine in western Finland, marking the occasion with a ceremonial event attended by invited guests and media.

New comment by philipkglass in "Last gasps of the rent seeking class?"

philipkglass — Fri, 27 Mar 2026 16:44:59 +0000

The small models that I can run at home are becoming more capable, and I have replaced some API-based tasks with local inference as they improve, but large open weights models are still a lot stronger. The nice thing with larger open weights models is that competing providers serve them at modest margins and prices. I don't have the hardware to run the largest Qwen models, but I can get API access at low cost. Since there are only modest barriers to new commercial inference providers for these models I'm not worried that API access to them will become drastically more expensive at some future time.

New comment by philipkglass in "Data centers are transitioning from AC to DC"

philipkglass — Wed, 25 Mar 2026 16:17:59 +0000

One more exception: uranium. It actually splits into smaller atoms when it's used as fuel.

New comment by philipkglass in "Blocking Internet Archive Won't Stop AI, but Will Erase Web's Historical Record"

philipkglass — Tue, 24 Mar 2026 18:44:38 +0000

I don't think that it is karma-gated, but you can only edit the comment for two hours after it's posted. Then it's frozen.

New comment by philipkglass in "The bridge to wealth is being pulled up with AI"

philipkglass — Tue, 24 Mar 2026 16:07:30 +0000

I think that the parent means Wall-E: https://en.wikipedia.org/wiki/WALL-E#Plot

New comment by philipkglass in "Nanopositioning Metrology, Gödel, and Bootstraps"

philipkglass — Mon, 23 Mar 2026 16:40:35 +0000

If you found this interesting and are a non-expert, I recommend browsing Foundations of Mechanical Accuracy by Wayne Moore (1970):

https://archive.org/details/Foundations_of_Mechanical_Accura...

Manufacturing got down to nanometer precision starting with macroscopic precision visible to the naked eyeball.

New comment by philipkglass in "FDA links raw cheese to outbreak; Makers "100% disagree," refuse recall"

philipkglass — Wed, 18 Mar 2026 15:27:00 +0000

A local small grocery chain started stocking raw milk (with many warnings) and I decided to risk consuming it to see for myself. I couldn't tell the difference between it and ordinary full fat milk. I wondered if it was a fraud (commercial milk falsely advertised as Local Forbidden Delicacy Milk), but maybe there's not much difference. Or maybe I am not a subtle taster. I also can't taste the superiority of an $80 bottle of wine when it's pitted against an $18 bottle.

New comment by philipkglass in "GPT‑5.4 Mini and Nano"

philipkglass — Tue, 17 Mar 2026 18:18:39 +0000

The crazy cheap models may be adequate for a task, and low cost matters with volume. I need to label millions of images to determine if they're sexually suggestive (this includes but is not limited to nudity). The Gemini 2.0 Flash Lite model is inexpensive and performs well. Gemini 2.5 Flash Lite is also good, but not noticeably better, and it costs more. When 2.0 gets retired this June my costs are going up.