I have often wondered how much a specialized local LLM could benefit an agentic tool like Gemini CLI. I would think there could be a good win for speed and minimizing token use if coding agents used a local model. A local model could handle a lot of the low level system interaction type tasks and then send the prompts that require deeper reasoning to frontier models. It seems wasteful and slow to use frontier models to figure out how to grep a codebase, run tests, git diff, etc.
Might Gemini CLI offload some of its prompts to FunctionGemma?
The most generic thing I can say is I really do like working at Google because its one of the few (maybe only) company that has models of all sizes and capabilities. Because of this research and product development is insanely fun and feels "magical" when things just click together.
Keep following the Google Developer channels/blogs whatever. Google as a whole is pushing hard in this space and I personally think is building stuff that felt like science fiction just 3 years ago.
HTML over the wire frameworks like HTMX, Hotwire (rails), LiveView (phoenix), Livewire (laravel), LiveView (django), etc. They all have the same basic idea with differences in how they achieve it. I feel this approach is overlooked, and it drives me crazy. There is a huge complexity cost attached with JS frontend app + backend that everyone seems to have accepted as reality. HTML over the wire (really need a catchy acronym maybe HotW) can greatly simplify and speed up development with pretty much the same end user experience as a react app.
I have lost track of the number of times I have heard that argument, and implemented the API only to never have other API clients, or the API clients that do come later have very different needs and I need to create another API for that. I don't think API re-use is a strong argument unless the other client will be consuming the API imminently. I've heard the of "course we will have a native mobile app" argument enough times to often identify it as wishful thinking. Many clients are often satisfied with with a good web app.
You can also reuse your backend if you server HTML, thanks to content negotiation. The frontend sends a header that indicates what type of content it accepts, and the backend serves the according type.
Also, mobile apps often have different API needs than webapps, so they end up getting different APIs anyway.
You can certainly try, but a lot of times the applications are different applications with different semantics so this doesn't work and you create a bunch more endpoints.
Hotwire Native would be a great option, but I will argue that I've never felt that I got JSON APIs for free with rails. Thousands of my rails dev hours spent writing APIs would say otherwise.
I realize terrestrial data centers have environmental risks, but are the risks greater for an orbital data center? I would think space debris, solar flares, or a bad actor satellite with a laser could do a lot of damage. Good luck repairing the orbital data center.
I can see how sharding could be difficult with a bigint FK, but UUIDv7 would still play nice, if I understand your point correctly. Monotonically increasing foreign keys have performance benefits over random UUIDv4 FKs in postgresql is the point of the article.
The undermining of science has given people like this more of a voice. US leadership has got to the point where science is largely disregarded and leaders just impose whatever they think is true regardless of facts.
I think it's more fundamental than that: science education has gotten so bad that people miss what should be pretty clear red flags.
The hydroxychloroquine debacle is one thing because there's no way as a layman to gauge that. It's still wrong, but I understand that people were worried and there's no intuitive route to "this won't help".
This feels like something where even laymen should be skeptical. Baseline, disinfectants are generally not medicine. Really, I would hope our education system was good enough that people hear "chlorinated disinfectant" and can jump to "probably an oxidizer and very bad for living things".
Hydroxychloroquine is definitely different from most of the bogus medical stuff going around.
When hydroxychloroquine was first proposed for COVID there was actually reason to think it might help. It is known to interfere with one of the mechanisms that COVID uses to enter cell membranes. The FDA in the US and equivalent regulators in many other countries gave it an emergency use authorization.
A few months later with more data it was found that it had some bad side effects and that it wasn't actually useful against COVID (most of the time COVID used a mechanism other than the one that hydroxychloroquine interfered with to enter cells).
Scientific method is practical. Scientific fact is a belief system, not unlike religion. This doesn’t undermine science, it’s just stating what these things are. Scientific belief can be helpful.
These silly beliefs though can be harmful as is the case with Chlorine Dioxide and that horse deworming medicine they said would cure Covid.
Don’t confuse or try to link these things together. The reason that the government is now full of idiots is that people voted those idiots in. It wasn’t due to clarity that science is part belief system.
I mean undermining of science in the sense of popular opinion. These politicians actively denigrate, de-fund, ignore science. Unfortunately, these same politicians were popular enough to gain power. They continue to attack science and treat science as if science and those who value it are their enemy.
Why do you think people voted these idiots in? In large part it's due to a distrust of science and rationality, "the system of the elites". Don't tell me that I need to vaccinate my kids! I'll cure my pancreatic cancer naturopathologically.
And the reason for believing these lies is because it's easier to listen to disgraced news outlets instead of reading a book or asking questions. It's a double edged sword: duped on one side and willful ignorance on the other.
Some might say, who cares about what idiots think? Well, these idiots get one vote, just like the rest of us.
> And the reason for believing these lies is because it's easier to listen to disgraced news outlets instead of reading a book or asking questions.
If it was just laziness, their belief wouldn't persist in the face of better information.
Willfully clasping onto an absurd premise this way strongly implies to me there is some sub-optimal psychology in play. There is a needful relationship inside that is being sustained (at a starvation level) by these exploitive external forces.
Your response, to me, is profound. I've lived in a deeply red voting state (OK) all my life. I never could understand how voters could vote against their best interests and to the harm of others. Perhaps your "sub-optimal psychology in play" is a reasonable answer.
You can certainly side with Descartes and believe in rationalism, or you can study David Hume and understand those beliefs for what they are.
You can still have and respect beliefs, but distinguishing fact from belief is purely in the realm of philosophical debate. That doesn’t mean you have to accept absence of absolute truth or facts, but there is much we don’t know and that we think we know but we don’t, and that is logical and illogical. Life is dichotomy.
No, you are witnessing democracy in action. People can disagree about "facts" and that is ok. People need to stop trying to impose things like this on others, have conversations with people they disagree with, and otherwise mind their business instead of advocating for a different group to impose on everyone whatever they think is true, disregarding individual autonomy and basic liberties.
I wish more of the web was like this. I miss the wild creativity of websites way back in the day. The web has mostly homogenized around what web UI should be. I love seeing weird experimental stuff made just for fun.
When I heard embedded postgres and sync, I immediately thought of pg's logical replication. ElectricSQL looks cool, but any chance of pg's native logical replication working with this?
Yes, we believe PostgreSQL's native logical replication is possible with PGlite. We have some ideas on how to achieve it, but we need more time to try them out.
Might Gemini CLI offload some of its prompts to FunctionGemma?