jvalencia

I've been writing my own, out of curiosity, with gemma4. I've been surprised how far I'm getting.

There's another problem here which is that there isn't enough content. I've on multiple occasions now thrown various news perspectives into AI and asked it to research what the actual facts of a contested issue were. In most cases, it comes down to one quote from one speech. The spin was pages and pages of commentary, most of which is opinion based. The news outlet wouldn't have enough to report if they just told you the quote.

Multimodal models are trained knowing how to understand encoded images. It really is magic. Base64 image data is a binary-to-text encoding scheme that represents image bytes as a printable ASCII string, formatted as data:[<mediatype>][;base64],<data>. We think of llms as only good at text, but any structured data is predictable. As long as it can be turned into an N dimensional vector that represents a complex idea in the LLM hidden weights, the output of the model treats that essentially like text. With sufficient training data, it understands the text in what to us looks like noise.

These sound like security nightmares.

How does security and isolation work? If someone else's account is compromised, how do I know I won't be? If instant is compromised, how do I know I won't be?

When a political party controls the science, you end up with say, Trump pushing one set of results, and Biden pushing another. It then becomes either pick the science that agrees with your politics, or throw up your hands in frustration. The average reader probably won't be able to dig into the fundamentals of the research and pull out the salient results, nor are they guaranteed it isn't policy pushed through overstated claims. It really undermines good science. It also falls back on the researchers who push science based on politics as well, so it isn't just the politicians.

Every few days I would log on. I was only 10-14 years old. It spread by word of mouth, and I just happened to have access to my dad's computer that had a modem. I'd hop on and play tradewars or similar. There were forums, mostly about hacking/pirating content. The forums were not too distant from what reddit feels like. As a young kid, it was also the only place where unfiltered information could be found, like how to make a bomb or how to get around copy protections. A lot of friends I had at the time where starting to do more serious file sharing, though the bandwidth kept that pretty limited.

Google for nonprofits is extremely generous. It's really not that bad in the end, and you only set it up once for a lot of benefit.

The trial and error was fueled by capitalism, trying to get the best product possible.

If it goes into a codified state system, it's regulated, resulting in a lack of motivation to take risks to make it better.

-ish. I often keep md files around and after a successful task. I ask Codex to write the important bits down. Then, when I come around to a similar task in the future, I have it start at the md file. It's like context that grows and is very localized. It helps when I'm going through multiple repos at multiple levels.

I think that's not wrong. Same principle, different sin... it looks like gambling, or the occult, or...

As a devout Baptist minister, this is likely about one of two things, avoiding the appearance of evil (gambling, 1 Thess 5:22 - Abstain from every form of evil), and giving up something for the sake of others (gambling addictions within the church, Rom 4:21 - or do anything that causes your brother to stumble).

The reality is that most churches recognize that they were too legalistic in the past, and so now address things like gambling more directly, and are perfectly ok with playing cards. FWIW YMMV :-)

Sure, but let's say you do EKS, you set it up once and then it's mostly done, including security, etc. You set up your own, then you upgrade every 6 months manually.... this is a cascading cost.

This takes time and effort, thus, lost opportunity cost. The thing that makes these providers worth it, is that it lets the business focus on their core competencies and just add-on as they scale without worrying about complexity. A business owner who hyper-optimizes for every contract is unlikely to be focusing on growing their business, even if their business is more efficient on paper.

Missed a link?

I found this pretty interesting, but would love to either see external inputs (news feed), or perhaps give the models more space. It seems like they might start collaborating more fully if they only had more space.

I'm curious if running them at slightly lower voltage would fix it or if it's a software thing.

or (3) get a battery

My first thought was why would someone halt their socials? Too much holiday time? Then I realized this was social media :-P. Socials to me are precisely NOT social media. Is this a common language usage now?

Large software projects are an interesting use case because once you get large is precisely when the framework becomes valuable.

A large enterprise project will need security, testing, auth, (AI now too). I'd hate to implement SAML without a library, that would be torture, and likely incompatible with most systems.

While I've often written small self projects from scratch, I wouldn't dream of building a large one that way unless you are sure to have an army of engineers and QA.

As an aside, this is where AI code fails as well. Speed of dev is easy, stability over time and compatibility is hard.

We have format commits so that we have separate non-logic commits that don't have to be aggravated over if we find files are all off.

California refineries are set to shut down and it has huge implications for the state: https://www.nytimes.com/2025/09/16/business/energy-environme...

It's like the invention of the washing machine. People didn't stop doing chores, they just do it more efficiently.

Coders won't stop being, they'll just do more, compete at higher levels. The losers are the ones who won't/can't adapt.

Well, and I think the future of LLMs is not just in the pure LLM, but the agentic ones. LLMs with deterministic tools to ferret out specifics. We're only starting here but the results will be far better than what we do today.

As someone who's sat on both sides of the table, getting tickets done is incredibly valuable. I don't really think an engineer is worth their salt if they can't do this. However, there's a transition that needs to happen when they hit mid-level where they need to be able to synthesize what the ticket is asking for and implement the right thing. No ticket or product manager is perfect, and awareness is often times a better driver of performance because certain tickets aren't done or are changed due to an engineer's competence.

From the flip side, management is looking at industry trends that the engineer simply doesn't see. It may be the current market is getting saturated with your product and the company really needs to pivot to remain competitive. No amount of further feature work or bug fixing will "fix" the market position. You have to do something different or you will lose sales. While fixing the existing product may make the existing customer happy, it won't continue to drive new revenue.

The only way to really make the two sides happy is to have a level of trust/communication that is rare. What engineer doesn't like to complain about management that keeps changing their mind? What manager doesn't like to complain about engineers that are out of touch with reality? Given this audience, I would say that if you're an engineer, there's an order to the skills: crush tickets, gain awareness of the product so you can do the "right" stuff, then solve your manager's market problems (not the product's).

Late 80s early 90s. And yes I played all the games you mentioned when the games came out. I especially loved simlife, but it didn't age, at all.

I did! And I enjoyed the mechanic, it was cute, incremental, and terribly fun.

I concede this may be true. First time I played D&D, I bought the rulebook and pored over it until I knew the rules and details -- and had a blast. Now, being a DM, no one reads the rules.

This is referencing a rule. I'm not sure I'm arguing against a clear reference, I like those. But I've certainly been at a table where we open the box, the host hasn't read the rules and frankly doesn't want to. I would hope you'd be excited about the board game, enjoy the book that comes with it, and then invite your friends over to play :-)

I recall my favorite part of opening up a new Nintendo game was poring over the manual. I didn't actually want it for instructions. It was the lore and the pictures and the dreams of what I could do with the game that made it fun.

Now, people are so media saturated that they're begging to remove as much content as possible in order to just get the clearest bit of information. I can't help we've lost something in society when we're not really enjoying or having fun unless we get immediate and overwhelming pleasure triggers.