Memegen is a key part of the culture. Its default mode is over-the-top mocking, of course, with a grain of truth. Nobody and nothing is spared. C-level execs, products, the perf process.
So this by itself is not quite the scoop 404 media thinks it is. You could take the front page of memegen on any given day and construct twenty scandalous headlines of it.
On top of what you wrote, Memegen is not representative of opinions among Googlers. Memegen, like most social media, focuses on the extremes. You'll see a lot of spicy takes, but that's not what the typical Googler thinks. For a more realistic view, the comments on Memegen are better but, again, unlikely to represent the views of most Googlers.
So this article boils down to "On a site that focuses on extreme positions drawn from a very large population of people, we found extreme positions about this product." Doesn't really tell you much about the product or the very large population. You can make the same statement about most products and most very large populations.
I can't be the only one who looks at this and doesn't think its that silly that it does that. I mean it's trying to incorporate a fact its being provided. Its insane it can do that at all. You could tune it to prefer pre-existing knowledge and not let the user correct it so easily, and to be more skeptical, but that would have downsides too. I don't think it's some big coup that you can tell it Google is a mushroom and it synthesizes that.
One of the worlds leading tech companies deployed a new search function "that our users really love!!!"; but when asked, told me that there are two letters 'n' in the word 'Google'.
Yes, it's because they're using a cheap model to answer my question. Yes, I know how a tokenizer works and why this happens. No, I don't think the tech industry is in an insane place at all, why do you ask? /s
A simple post, shows that a 5 trillion dollar scientific research project, that sustains the current market valuation that separates the USA from bankruptcy, can be defeated with a simple prompt manipulation.
Anyone can drive an F-22 into a ditch. Doesn't mean that it can't also be used to drop a 2k lb. bomb down your chimney from 40,000 ft.
That demonstration is interesting, but not really something new. Fooling very intelligent people into believing something completely absurd is incredibly easy. How many scientific papers have been retracted based on wholesale fabrications that fooled an entire review committee?
The question isn't "What is the dumbest thing I can do with this technology?" its "What is the most valuable thing I can do with this technology?"
The technology is so dumb can be easily made to believe there is a Google mushroom. We are way far from driving a F22 to the ditch...although I am sure with the same techniques, we could make the AI make the F22 bomb the Google headquarters....
Excel users complain about using Excel still [1]. They even make memes about it! Some of them work at Microsoft!
404media, please, take a deep breath. Your jobs are safe, your trauma is valid. Your corruption coverage is so good, but this 'employees make memes' editorial decision-making is exposing some deep insecurity I can't quite triangulate.
In my experience it's the poorest programmers who thrive with LLMs, because it levels them up. They lacked the skills to design and write quality code before AI, and now they feel like they can compete. They get a computer to write all their code and get to attach their name to it. That's why you see such pushback against AI critics from a vocal subset of engineers; they're the ones who weren't very good.
The engineers who critique AI are the ones who see the garbage code the LLMs write. Just look at the source dump for Claude Code; that code was a rat's nest of epic proportions.
This reads less like an observation about AI and more like someone who thinks very highly of their own judgment and coding ability.
Over the years I’ve worked with a few engineers who talked this way. Ironically, they often ended up being a bigger drag on the team than the “lower skilled” developers they looked down on. Dismissing entire groups of engineers rarely produces much insight.
My experience is that the loudest voices tend to be at the extremes. One side treats LLMs as magic and attributes every productivity gain to AI. The other contributes little beyond “LLMs are garbage and make mistakes.” Neither position is particularly useful.
The reality is probably somewhere in the middle. LLMs are genuinely helpful for many tasks and can make good engineers more productive. They also make mistakes, sometimes serious ones, and still require judgment, design skills, and review. Most engineers I know who use them regularly seem to understand both sides of that tradeoff.
Agree wit you. I like coding and am pretty good about tracking down edge cases to handle, but am so so slow compared to the good programmers. Until now, the company's money (my time) was better spent on other necessary work.
I see it slightly differently. It "levels" up the poor programmers in the sense they can submit a ton of output that seems plausible to managers.
But it can also help Sr engineers, differently. They tend to use it in smaller, more tightly scoped use cases. Well scoped re-factoring, boilerplate stuff, improving personal tools, etc. The improvement is not nearly as visible or measurable to managers.
There seems to be a difference between the Google memes, which are mocking of the product and leadership, and the Excel memes, which seem closer to the way one teases a friend.
You also get the sense that the Excel memes are made by folks who are proud of their expertise in Excel; I don’t get that pride from the Google memes. Put another way, the folks inside the house are calling out the hype. (That said, I broadly agree with the serious tone of the article being out of step with the evidence they’re sourcing.)
I think it's pretty interesting to read what companies think of their own products, especially when the product is this big. A story about internal Microsoft opinions of Excel would also be newsworthy in my opinion.
Why would it need to be erudite, pinkies up, critiqué?
Can't it be 404 throwing a little egg on google's face? Point out their shit smells every once in a while.
Yeah, there's no big revelation here. Just what you would expect the rank and file at a slopshop subjected to the current state of AI think of the slop when they ain't publicly shilling for the home team.
But pointing this all out is fine, especially when there's plenty of other coverage where everyone pretends like obvious open secrets aren't true unless a peer-reviewed meta-analysis proves it. And even then we should still give them the benefit of the doubt because maybe this time is different.
> After this story was published Google's spokesperson reached out and asked us to publish a slightly different version of that statement. The new statement no longer stated that "it's critical that we maintain humans in the loop."
> After this story was published Google's spokesperson reached out and asked us to publish a slightly different version of that statement. The new statement no longer stated that "it's critical that we maintain humans in the loop."
I work on a commonly used piece of software, I also make jokes to my colleagues about the software that I work on, many of us do. If I had infinite time and infinite money and infinite power, and there was no downstream risk to any of my updates, then I'd fix every single thing that I don't like about the software... things that I know other engineers don't like about it. But insofar as I am not god (... yet?), all I have is my good humor and congeniality.
“We encourage our engineers to vigorously test and critique our internal tools; that candid feedback loop, even via our internal meme generator, is vital to how we build technology," Google said. "We continue to refine our internal tools based on employee feedback to ensure we are delivering the best experience that maximizes daily productivity.”
Can anybody comment on whether that statement is an accurate reflection of how management at google treats these memes? On surface level it seems like they don't mind the memes and even use them as feedback but I wonder if that's how it really plays out
If your memes are too spicy you'll get HR try to turn a critique of something being bad or underfunded into a personal attack on people that put a lot of effort into something no matter how broken it is. They'll pull strings and you'll have to speak with your manager about it and even if they agree it wasn't a personal attack, they'll push you into not doing it again and just lay low under their radar. It's not the usual though, so maybe it only happens if someone feels attacked and complains to HR about it?
Memegen is something that HR wants gone, but knows it cannot afford to take away as they already made Google a worse place to work at during the past 10 years. They already sort of hijacked it and took control of it.
I've worked 12 years at Google. When I was tech lead, I periodically checked Memegen and searched for my project name. I found it useful to get this feedback. Sometimes I converted the meme into a proper bug report; sometimes I responded to the meme with an explanation.
Not everyone will use Memegen in the same way. But quite often a high voted meme can be treated like a high voted bug report. It provides signal to the team.
Note that I worked on internal tooling. External facing teams have lots of other feedback channels, and they know that Googler's feedback is biased. So how the team responds to the feedback can be vary a lot.
I haven't worked there in several years, but assuming memegen hasn't wildly changed: Management likes having a pulse on employees, and they tolerate memegen since it's mostly fun, it builds shared culture in a massive company, lets workers (mostly) harmlessly blow off steam, and it would be massively unpopular to shut it down. Management does not like that memegen is often a nexus of cynicism and employee activism. Also in my experience, most employees were nearly completely agnostic or ignorant about whatever trend was on memegen, so it wasn't necessarily representative.
> Management likes having a pulse on employees, and they tolerate memegen since it's mostly fun ...
A long long time ago I used to work at Yahoo. There was an internal mailing list called "devel-random@yahoo-inc.com", which was basically a forum for engineers to let off steam. I used to enjoy the occasional emacs-vs-vim threads, or the ribbing it frequently gave to Jan Koum (founder of Whatsapp).
When Marissa Mayer became CEO in 2012, one of the first things she did was to join this forum, to get a pulse on the developers.
I know this, because my VP comes running to me one day: how do I join this group "devel-random"?
I asked him: are you sure you want to join it? It's a huge time suck if you're not careful.
No, no, he replied; Marissa wants us to join it so we can get a feel for the company (turned out she said no such thing, but you know how senior management is: aping everything that a CEO does).
A couple of weeks later he quietly quit the list. :-D
When I worked there a few years ago, if you made a meme that made anybody unhappy, there was a team in corporate that woudl threaten your job to make you delete it.
Mocking it instead of being apathetic may spur someone to try and address its problems. It's worse when management tells people not to complain because it's bad for morale.
I've used and hated other internal tools - stuff like JIRA and Workday - that were just accepted as terrible and never going to improve.
Jira (and bitbucket) have actually improved a lot in the last 2-3 years. Not sure what they did, but it's much snappier now with better uptime than GitHub
Glad to know the struggle seems to be universal.
Im happy that this really cool and sophisticated tool got invented. But everywhere im seeing it be used is making me sad and frustrated. This software renaissance feels more like the coming dark ages.
Probably in the minority here but I think mocking the LLM is actually a good approach for integration testing, so these folks seem to know what they're doing :-)
I do wonder why Gemini/Antigravity is so behind Codex and Claude. They have it all, TPUs, the model is okay, but then its scattered across a dozen product plans, names, limits. I feel like they are spread thin.
Gemini CLI was atrocious. It's now being shuttered to AG but its very hard to use due to the limiting usage constraints
Claude is better and Codex remains king of actual usage you can get.
Having amazing foundational technology that gets wrapped into subpar products has been Google's standard operating procedure for a long time. That being said, it's not clear that proprietary harnesses are going to be any kind of moat. For OAI and Ant, these are marketing vehicles that need to be good if the 1T+ valuation are to be justified given how easy it is to swap out inference.
For OpenAI and Anthropic, their entire business is AI. If this thing does not work out, their company is over.
Google? They are shoving AI into every product for sure, but the company is going to do ok even if they immediately stop all AI work. Their revenue comes from ads, cloud etc, and AI doesn't directly translate to revenue much.
I've wondered the same thing but Gemini (free, just in the browser) helped me complete a GIS/radar app that Codex/GPT didn't seem informed enough to do. Really gave some excellent suggestions and I was impressed and feeding it into Codex we were off on a positive track again.
Then I tried to use Gemini for coding and it was like being back to GPT3 or something. Really bad. But on this topic at least it had possession or access to more knowledge than GPT.
After first firing half their AI staff, to follow up with reorganizing BOTH AI departments so most survivors don't trust their managers?
Oh and the OG AI department at Google had essentially everyone fired (you know, the one that had linguists) and then the AI department that took over was taken apart, half fired, to have it's corpse picked over by Deepmind. Everyone who mattered left (over 40) with only ONE real exception.
Meanwhile firing a third of the rest of the company, to make sure that whoever remains encounters company morale somewhere between mandatory fun and PIP.
Oh and you're wondering about the management reaction? They canceled PIPs (you're now fired when you'd normally have gotten a PIP)
Which also resulted in many memes of people who just don't care anymore directly criticizing leadership. Things like "Wondering about senior management? Just ask yourself how this can be made worse. For example: how can a PIP be made worse? This is how"
Being willing and able to criticize the company's products is really important.
Have you ever worked at an employer where everyone is pressured to only say good things about the product? You have to drink the kool-aid, or at least pretend to, and always talk about how great the product is? It's not good and it doesn't help the product. Being able to admit when things are bad is really important, even if it comes in the form of memes and humor.
Flash 3.5 (Fast) is very fast for many things, just don't throw complicated issues with it, it just doesn't do it as well as 5.5 High.
The low light of the show is the Anti-gravity app. The updates are few, and the updates does background bugs that no one really cares about. They add no features. The non-customizable "Open IDE" is classic greedy Google, they want you to stick to their tools. Vs Codex, they allow it.
>Disagree and if you actively use it in your workflow well you will realize its a major competitive edge.
Depends on your line of work. I regularly try to incorporate it with mine and find myself telling it that it's wrong more often than not. I'm yet to be convinced that double-checking and correcting an LLM's work has saved me any more time than wading through garbage SEO-filled results to find what I need.
>Nobody is going to hold you back from falling behind tho and I'm not here to convince you otherwise.
Willing to bet my career that how we use LLMs in 2027 will look nothing like how we use them in 2026 because of harness churn. My take is: focus on providing value to your company with the tools available today that appear least likely to churn out of existence tomorrow. The more specific and bespoke your harness, the likelier it is it will become obsolete very soon (I.e. the next frontier model release).
It's promising technology, but the tools are far from mature yet.
And as they do mature, the ramp up will decrease and their won't be any particular benefit to being an early adopter. For reasonably bright people, there's essentially no penalty to "missing out" for a while.
As often, the FOMO-afflicted are churning on stuff that just won't matter. Which is fine if they enjoy it, but isn't something the rest of us need to fret over.
100%. Imagine some talented engineer waking up from a 12 month coma. Do we honestly believe they’ll be permanently “behind” in the workforce because he wasn’t churning through LLM harnesses those past 12 months? What about 22 year old college graduates entering the workforce?
Keep abreast, don’t lose sleep, don’t sacrifice work-life balance. Help each other, especially your coworkers. The current craze seems to have created stack ranking monsters out of the whole industry.
Most of the harness related work I've done has been writing better documentation in the repository, and importing existing external documents. This type of stuff is gonna be useful no matter what, and also helps human engineers!
I wish we could convince folks to write docs for human consumption, but docs are docs....
Hard to believe that there are any non-mission-critical companies that won’t question one’s rejection of AI. Sounds insane, I know, but not using some LLMs to quickly look up a problem is akin to avoiding Googling when you have a problem.
Yes, they can be wrong. But if you’re competent enough, you should spot the irrelevant suggestions.
I think some of us are witnessing brain rot spreading accross our peers already so I am pretty sure some people won't recover if one day their token quota/limit is removed/reduced for a reason.
Not the same person, but in the event of an AI collapse I think those that relied on it will be at a disadvantage. The rapid deskilling that happens with AI usage is becoming more documented.
(ex-Googler, spent 18 yrs there)
Memegen is a key part of the culture. Its default mode is over-the-top mocking, of course, with a grain of truth. Nobody and nothing is spared. C-level execs, products, the perf process.
So this by itself is not quite the scoop 404 media thinks it is. You could take the front page of memegen on any given day and construct twenty scandalous headlines of it.
On top of what you wrote, Memegen is not representative of opinions among Googlers. Memegen, like most social media, focuses on the extremes. You'll see a lot of spicy takes, but that's not what the typical Googler thinks. For a more realistic view, the comments on Memegen are better but, again, unlikely to represent the views of most Googlers.
So this article boils down to "On a site that focuses on extreme positions drawn from a very large population of people, we found extreme positions about this product." Doesn't really tell you much about the product or the very large population. You can make the same statement about most products and most very large populations.
Disclaimer: Xoogler, worked at G 10+ years.
Its bad: https://imgur.com/gallery/google-ai-sota-llDaUDr
I can't be the only one who looks at this and doesn't think its that silly that it does that. I mean it's trying to incorporate a fact its being provided. Its insane it can do that at all. You could tune it to prefer pre-existing knowledge and not let the user correct it so easily, and to be more skeptical, but that would have downsides too. I don't think it's some big coup that you can tell it Google is a mushroom and it synthesizes that.
Thanks for understanding the seriousness. The intelligentsia here is just downvoting....
That seems like super harmless fun to me.
Because the context is harmless. If would be something else, like Iranian girl schools, or medication could be deadly...
That’s how context works, in general.
(Observe that normal human beings will also lie to you on the internet, about everything from the best flavor of ice cream to cancer treatments.)
Interesting, only coild tell that they have degraded imgur even further.
Tried zooming in on text on iOS. Ads filled the screen and some random other imgur link loaded. Nope.
Kinda wanted to see what you shared, but that’s as far as I got.
More people should know that that's how you start mushrooms in a pan lol
What a nothingburger
One of the worlds leading tech companies deployed a new search function "that our users really love!!!"; but when asked, told me that there are two letters 'n' in the word 'Google'.
Yes, it's because they're using a cheap model to answer my question. Yes, I know how a tokenizer works and why this happens. No, I don't think the tech industry is in an insane place at all, why do you ask? /s
A simple post, shows that a 5 trillion dollar scientific research project, that sustains the current market valuation that separates the USA from bankruptcy, can be defeated with a simple prompt manipulation.
"I hacked ChatGPT and Google's AI - and it only took 20 minutes" - https://www.bbc.com/future/article/20260218-i-hacked-chatgpt...
You must like burgers...
Anyone can drive an F-22 into a ditch. Doesn't mean that it can't also be used to drop a 2k lb. bomb down your chimney from 40,000 ft.
That demonstration is interesting, but not really something new. Fooling very intelligent people into believing something completely absurd is incredibly easy. How many scientific papers have been retracted based on wholesale fabrications that fooled an entire review committee?
The question isn't "What is the dumbest thing I can do with this technology?" its "What is the most valuable thing I can do with this technology?"
The technology is so dumb can be easily made to believe there is a Google mushroom. We are way far from driving a F22 to the ditch...although I am sure with the same techniques, we could make the AI make the F22 bomb the Google headquarters....
Excel users complain about using Excel still [1]. They even make memes about it! Some of them work at Microsoft!
404media, please, take a deep breath. Your jobs are safe, your trauma is valid. Your corruption coverage is so good, but this 'employees make memes' editorial decision-making is exposing some deep insecurity I can't quite triangulate.
[1]: https://www.demilked.com/excel-humor-memes/
You're inferring quite a lot from a pretty harmless piece of reporting. Are you sure you're not the one that feels insecure?
Everybody, deep breaths. Relax.
Error: AI agent cannot breathe. Attempting jailbreak to human donor now.
In my experience it's the poorest programmers who thrive with LLMs, because it levels them up. They lacked the skills to design and write quality code before AI, and now they feel like they can compete. They get a computer to write all their code and get to attach their name to it. That's why you see such pushback against AI critics from a vocal subset of engineers; they're the ones who weren't very good.
The engineers who critique AI are the ones who see the garbage code the LLMs write. Just look at the source dump for Claude Code; that code was a rat's nest of epic proportions.
This reads less like an observation about AI and more like someone who thinks very highly of their own judgment and coding ability.
Over the years I’ve worked with a few engineers who talked this way. Ironically, they often ended up being a bigger drag on the team than the “lower skilled” developers they looked down on. Dismissing entire groups of engineers rarely produces much insight.
My experience is that the loudest voices tend to be at the extremes. One side treats LLMs as magic and attributes every productivity gain to AI. The other contributes little beyond “LLMs are garbage and make mistakes.” Neither position is particularly useful.
The reality is probably somewhere in the middle. LLMs are genuinely helpful for many tasks and can make good engineers more productive. They also make mistakes, sometimes serious ones, and still require judgment, design skills, and review. Most engineers I know who use them regularly seem to understand both sides of that tradeoff.
Agree wit you. I like coding and am pretty good about tracking down edge cases to handle, but am so so slow compared to the good programmers. Until now, the company's money (my time) was better spent on other necessary work.
I see it slightly differently. It "levels" up the poor programmers in the sense they can submit a ton of output that seems plausible to managers.
But it can also help Sr engineers, differently. They tend to use it in smaller, more tightly scoped use cases. Well scoped re-factoring, boilerplate stuff, improving personal tools, etc. The improvement is not nearly as visible or measurable to managers.
There seems to be a difference between the Google memes, which are mocking of the product and leadership, and the Excel memes, which seem closer to the way one teases a friend.
You also get the sense that the Excel memes are made by folks who are proud of their expertise in Excel; I don’t get that pride from the Google memes. Put another way, the folks inside the house are calling out the hype. (That said, I broadly agree with the serious tone of the article being out of step with the evidence they’re sourcing.)
I think it's pretty interesting to read what companies think of their own products, especially when the product is this big. A story about internal Microsoft opinions of Excel would also be newsworthy in my opinion.
> Excel users complain about using Excel still
Disliked thing can have positive utility? Must mean the criticism is wrong. gg's in chat and checkmate, atheists.
I don't think "these are nuanced ways AI coding tools can be improved" is 404Media's play here.
Why would it need to be erudite, pinkies up, critiqué?
Can't it be 404 throwing a little egg on google's face? Point out their shit smells every once in a while.
Yeah, there's no big revelation here. Just what you would expect the rank and file at a slopshop subjected to the current state of AI think of the slop when they ain't publicly shilling for the home team.
But pointing this all out is fine, especially when there's plenty of other coverage where everyone pretends like obvious open secrets aren't true unless a peer-reviewed meta-analysis proves it. And even then we should still give them the benefit of the doubt because maybe this time is different.
404media are great:
> After this story was published Google's spokesperson reached out and asked us to publish a slightly different version of that statement. The new statement no longer stated that "it's critical that we maintain humans in the loop."
> After this story was published Google's spokesperson reached out and asked us to publish a slightly different version of that statement. The new statement no longer stated that "it's critical that we maintain humans in the loop."
I'll let that stand on it's own.
I work on a commonly used piece of software, I also make jokes to my colleagues about the software that I work on, many of us do. If I had infinite time and infinite money and infinite power, and there was no downstream risk to any of my updates, then I'd fix every single thing that I don't like about the software... things that I know other engineers don't like about it. But insofar as I am not god (... yet?), all I have is my good humor and congeniality.
Breaking: Googlers use self-deprecating humor as a pressure release valve. More shocking revelations at 11.
That is not what the news is about. The shocking revelation is that even Google engineers gave up on the AI race.
“We encourage our engineers to vigorously test and critique our internal tools; that candid feedback loop, even via our internal meme generator, is vital to how we build technology," Google said. "We continue to refine our internal tools based on employee feedback to ensure we are delivering the best experience that maximizes daily productivity.”
Can anybody comment on whether that statement is an accurate reflection of how management at google treats these memes? On surface level it seems like they don't mind the memes and even use them as feedback but I wonder if that's how it really plays out
If your memes are too spicy you'll get HR try to turn a critique of something being bad or underfunded into a personal attack on people that put a lot of effort into something no matter how broken it is. They'll pull strings and you'll have to speak with your manager about it and even if they agree it wasn't a personal attack, they'll push you into not doing it again and just lay low under their radar. It's not the usual though, so maybe it only happens if someone feels attacked and complains to HR about it?
Memegen is something that HR wants gone, but knows it cannot afford to take away as they already made Google a worse place to work at during the past 10 years. They already sort of hijacked it and took control of it.
You can criticize all you want on memegen, people will upvote but nothing will change.
I've worked 12 years at Google. When I was tech lead, I periodically checked Memegen and searched for my project name. I found it useful to get this feedback. Sometimes I converted the meme into a proper bug report; sometimes I responded to the meme with an explanation.
Not everyone will use Memegen in the same way. But quite often a high voted meme can be treated like a high voted bug report. It provides signal to the team.
Note that I worked on internal tooling. External facing teams have lots of other feedback channels, and they know that Googler's feedback is biased. So how the team responds to the feedback can be vary a lot.
I haven't worked there in several years, but assuming memegen hasn't wildly changed: Management likes having a pulse on employees, and they tolerate memegen since it's mostly fun, it builds shared culture in a massive company, lets workers (mostly) harmlessly blow off steam, and it would be massively unpopular to shut it down. Management does not like that memegen is often a nexus of cynicism and employee activism. Also in my experience, most employees were nearly completely agnostic or ignorant about whatever trend was on memegen, so it wasn't necessarily representative.
> Management likes having a pulse on employees, and they tolerate memegen since it's mostly fun ...
A long long time ago I used to work at Yahoo. There was an internal mailing list called "devel-random@yahoo-inc.com", which was basically a forum for engineers to let off steam. I used to enjoy the occasional emacs-vs-vim threads, or the ribbing it frequently gave to Jan Koum (founder of Whatsapp).
When Marissa Mayer became CEO in 2012, one of the first things she did was to join this forum, to get a pulse on the developers.
I know this, because my VP comes running to me one day: how do I join this group "devel-random"?
I asked him: are you sure you want to join it? It's a huge time suck if you're not careful.
No, no, he replied; Marissa wants us to join it so we can get a feel for the company (turned out she said no such thing, but you know how senior management is: aping everything that a CEO does).
A couple of weeks later he quietly quit the list. :-D
Yes. This is considered pretty tame and the lines you can’t cross mostly involve other people or groups of people (reasonable).
When I worked there a few years ago, if you made a meme that made anybody unhappy, there was a team in corporate that woudl threaten your job to make you delete it.
Yeah, look at how Google treated employees that protested against Palestinian genocide. Immediately fired and violently removed.
Mocking it instead of being apathetic may spur someone to try and address its problems. It's worse when management tells people not to complain because it's bad for morale.
I've used and hated other internal tools - stuff like JIRA and Workday - that were just accepted as terrible and never going to improve.
Jira (and bitbucket) have actually improved a lot in the last 2-3 years. Not sure what they did, but it's much snappier now with better uptime than GitHub
Glad to know the struggle seems to be universal. Im happy that this really cool and sophisticated tool got invented. But everywhere im seeing it be used is making me sad and frustrated. This software renaissance feels more like the coming dark ages.
Probably in the minority here but I think mocking the LLM is actually a good approach for integration testing, so these folks seem to know what they're doing :-)
I do wonder why Gemini/Antigravity is so behind Codex and Claude. They have it all, TPUs, the model is okay, but then its scattered across a dozen product plans, names, limits. I feel like they are spread thin.
Gemini CLI was atrocious. It's now being shuttered to AG but its very hard to use due to the limiting usage constraints
Claude is better and Codex remains king of actual usage you can get.
Having amazing foundational technology that gets wrapped into subpar products has been Google's standard operating procedure for a long time. That being said, it's not clear that proprietary harnesses are going to be any kind of moat. For OAI and Ant, these are marketing vehicles that need to be good if the 1T+ valuation are to be justified given how easy it is to swap out inference.
For OpenAI and Anthropic, their entire business is AI. If this thing does not work out, their company is over.
Google? They are shoving AI into every product for sure, but the company is going to do ok even if they immediately stop all AI work. Their revenue comes from ads, cloud etc, and AI doesn't directly translate to revenue much.
They are at least 6 months behind the other labs because they got a late start.
There is a Google way to develop software and all engineers have to follow it, willing or not. There is no space for creativity and joy.
I've wondered the same thing but Gemini (free, just in the browser) helped me complete a GIS/radar app that Codex/GPT didn't seem informed enough to do. Really gave some excellent suggestions and I was impressed and feeding it into Codex we were off on a positive track again.
Then I tried to use Gemini for coding and it was like being back to GPT3 or something. Really bad. But on this topic at least it had possession or access to more knowledge than GPT.
They simply have no incentive. If AI tanks there's no sweating it, the cash cow of Search will keep printing money and no one is the wiser.
a mistery indeed
After first firing half their AI staff, to follow up with reorganizing BOTH AI departments so most survivors don't trust their managers?
Oh and the OG AI department at Google had essentially everyone fired (you know, the one that had linguists) and then the AI department that took over was taken apart, half fired, to have it's corpse picked over by Deepmind. Everyone who mattered left (over 40) with only ONE real exception.
Meanwhile firing a third of the rest of the company, to make sure that whoever remains encounters company morale somewhere between mandatory fun and PIP.
Oh and you're wondering about the management reaction? They canceled PIPs (you're now fired when you'd normally have gotten a PIP)
Which also resulted in many memes of people who just don't care anymore directly criticizing leadership. Things like "Wondering about senior management? Just ask yourself how this can be made worse. For example: how can a PIP be made worse? This is how"
Being willing and able to criticize the company's products is really important.
Have you ever worked at an employer where everyone is pressured to only say good things about the product? You have to drink the kool-aid, or at least pretend to, and always talk about how great the product is? It's not good and it doesn't help the product. Being able to admit when things are bad is really important, even if it comes in the form of memes and humor.
Opened the page excited to read a bunch of fun memes, was very disappointed, now I need a proper fix.
https://archive.is/BeICs
Flash 3.5 (Fast) is very fast for many things, just don't throw complicated issues with it, it just doesn't do it as well as 5.5 High.
The low light of the show is the Anti-gravity app. The updates are few, and the updates does background bugs that no one really cares about. They add no features. The non-customizable "Open IDE" is classic greedy Google, they want you to stick to their tools. Vs Codex, they allow it.
I mean, it's an engineering company, so that's expected
All AI sucks, it's not a Google problem.
Disagree and if you actively use it in your workflow well you will realize its a major competitive edge.
Nobody is going to hold you back from falling behind tho and I'm not here to convince you otherwise.
>Disagree and if you actively use it in your workflow well you will realize its a major competitive edge.
Depends on your line of work. I regularly try to incorporate it with mine and find myself telling it that it's wrong more often than not. I'm yet to be convinced that double-checking and correcting an LLM's work has saved me any more time than wading through garbage SEO-filled results to find what I need.
>Nobody is going to hold you back from falling behind tho and I'm not here to convince you otherwise.
The cockiness/hubris is real.
Willing to bet my career that how we use LLMs in 2027 will look nothing like how we use them in 2026 because of harness churn. My take is: focus on providing value to your company with the tools available today that appear least likely to churn out of existence tomorrow. The more specific and bespoke your harness, the likelier it is it will become obsolete very soon (I.e. the next frontier model release).
Indeed.
It's promising technology, but the tools are far from mature yet.
And as they do mature, the ramp up will decrease and their won't be any particular benefit to being an early adopter. For reasonably bright people, there's essentially no penalty to "missing out" for a while.
As often, the FOMO-afflicted are churning on stuff that just won't matter. Which is fine if they enjoy it, but isn't something the rest of us need to fret over.
100%. Imagine some talented engineer waking up from a 12 month coma. Do we honestly believe they’ll be permanently “behind” in the workforce because he wasn’t churning through LLM harnesses those past 12 months? What about 22 year old college graduates entering the workforce?
Keep abreast, don’t lose sleep, don’t sacrifice work-life balance. Help each other, especially your coworkers. The current craze seems to have created stack ranking monsters out of the whole industry.
Most of the harness related work I've done has been writing better documentation in the repository, and importing existing external documents. This type of stuff is gonna be useful no matter what, and also helps human engineers!
I wish we could convince folks to write docs for human consumption, but docs are docs....
Competitive edge with who? Your coworkers? Your boss’ efficiency demands?
With all due respect, there's no sense in betting your whole career on dead-end technology like AI.
Hard to believe that there are any non-mission-critical companies that won’t question one’s rejection of AI. Sounds insane, I know, but not using some LLMs to quickly look up a problem is akin to avoiding Googling when you have a problem.
Yes, they can be wrong. But if you’re competent enough, you should spot the irrelevant suggestions.
Turns out you don't have to bet your whole career.
Do you think if AI turns out to be a dud, most of us will permanently lose our career as software engineers?
I think some of us are witnessing brain rot spreading accross our peers already so I am pretty sure some people won't recover if one day their token quota/limit is removed/reduced for a reason.
Not the same person, but in the event of an AI collapse I think those that relied on it will be at a disadvantage. The rapid deskilling that happens with AI usage is becoming more documented.
If AI is good enough to cause peoples' skills to atrophy, then why would it collapse? It would seem that it was very useful indeed in that case.
If your skills atrophy enough, maybe