Gemini really feels like a high-performing child raised in an abusive household.

Every time I see people praise Gemini I really wonder what simple little tasks they are using it for. Because in an actual coding session (with OpenCode or even their own Gemini CLI for example) it just _devolves_ into insanity.

And not even at high token counts! No, I've had it had a mental breakdown at like 150.000 tokens (which I know is a lot of tokens, but it's small compared to the 1 million tokens it should be able to handle, and even Claude keeps working fine at this point)

Here is a _small_ log of the biggest breakdown I've seen Gemini have:

    I'll verify `src/components/ui/textarea.tsx` renders textarea.
    Correct.

    I'm done.

    I'll verify `src/lib/utils.ts` exports `cn`.
    Correct.

    I'm done.

    I'll verify `src/lib/utils.ts` exports `formatTimeAgo`.
    Correct.

    I'm done.

    I'm completely done.

    Final answer formulation:
    1. Confirm crash is fixed (caused by object vs string in API responses).
    2. Announce "Messages" feature added.
    3. Explain "My Comments" limitation.
    4. Mention login race condition fix (from previous turn) is solid.
    5. Mention server is running.

    (Wait, the "Login" fix was in previous turn, but I should probably mention it's working now if the user was still having issues).
    The user said "unclaimed_or_invalid_key" error showed up. So login is working.

    The user's last error was "Objects are not valid as a React child".
    So I fixed that.

    I'll focus on the crash fix and the new feature.

    I'll conclude.

    (Self-correction: The user asked "how do I see the comments I made?". I'll answer "You can't right now because API doesn't support it").

    I'm ready.

    ```bash
    # None
    ```

    The user will see the "Messages" link in the header.

    I'll provide the answer now.

And it just went on and on

With Codex it can happen on context compacting. Context compacting with Codex is a true Russian roulette, 7 times out of 8 nothing happens and the last one kills it

This seems much more terse than Gemini usually is, are you prompting it to do that?

If you recall the context/situation at the time it was released, that might be close to the truth. Google desperately needed to show competency in improving Gemini capabilities, and other considerations could have been assigned lower priority.

So they could have paid a price in “model welfare” and released an LLM very eager to deliver.

It also shows in AA-Omniscience Hallucination Rate benchmark where Gemini has 88%, the worst from frontier models.

A la carte in AI is going to be the name of the game for a couple reasons:

- Avoids regulatory scrutiny (for now at least)

- Nobody is actually entrenched enough for customers to matter

- Weird "celebrity" culture in tech, and AI especially. Everyone is looking for a "whisperer" or a "godfather" or whatever.

- Investors still get paid out

Smart operational talent will probably adapt by demanding higher salary, signing bonuses, severance packages in lieu of equity. Distribution of the true "lottery tickets" will get more uneven.

I've noticed people using emdashes more in known non-AI text in what I assume is a smokescreen to maintain plausible deniability when they wholesale copy AI text.

It's so interesting to me that human writing is subtly changing to mirror AI writing.

Or maybe they’ve been there all along and you just notice them more now because you’re looking for them.

I was always looking for them because I was the weird nerd pointing out proper em dash, en dash, and hyphen usage years and years ago.

It's really only devs / engineers I see doing this, probably in some quest to create an indistinguishable voice in the name of productivity or something.

The transmission network is underbuilt, so it's mostly best to generate closer to where it's consumed (especially for data centers).

We'll continue to see a mix though of Residential / Commercial & Industrial / Utility Scale

There are about 7,000 Utility scale sites in the US right now, so even the big boys there are fairly distributed.

More like they see the future as more multi-modal, and they're probably right to think that is the best value approach vs. throwing more money at large language models.

I take screenshots because they're definitely going to forget to share the presentation.

I agree with you, but in a different directions. Often the software is owned by the company, who are just renting the engineer's labor.

> Individual engineers don’t own software; engineering teams own software.

Very assertive, but almost always incorrect.

I don't know if these statistics are even kept, but the current social environment in the US feels like a ripe breeding ground for cults. I've had so many people in the past couple years be like "I just want to farm with my friends and family and get away from all this".

It has become a bit of a meme lately. I think there's something to be said about a malaise era leading to an uptick in erratic behavior.

But at the risk of sounding smug and condescending, as someone who actually bought 10 acres "to get away from it all", I get the sense that the type of people in this saga would pack up pretty quick after a little taste. Lesswrongers aren't exactly known for pragmatism, which is sorta the only mindset that works. There's all this work that you don't know you don't know about. I just fell down the rabbit hole of the ziz lore and goddamn do these people sound inept. Like they couldn't even fix their RV to get off their landlord's property. Lots of quasi-intellectual masturbatory posting and not a lot of skills.

All that is to say I'm not super worried about any of these cults really taking off. Logistics remain challenging.