> I think most of it is bot driven spam by the various labs. It's hard not to notice 3 month old accounts with very strong opinions about various frontier labs and little else.
I've assumed the same as well.
I also assume that many of the companies developing these models engage in benchmaxxing.
At my company we've developed our own internal benchmarks for evaluating LLM models as they become available. The benchmarks are tailored to our particular use cases but the utility and knowledge our benchmarks assess is still fairly universally applicable. I see wide differences between what our internal benchmarks report and what the major benchmarks do.
There was a whole lot of fanfare about how amazing GLM 5.2 was when it was released, but it was pure rubbish on our internal benchmark -- far behind OpenAI, Anthropic, Gemini, DeepSeek, etc. I don't know how to reconcile the fact that GLM 5.2 performed very well on some of the major public benchmarks, but consistently performs so poorly on ours. OpenAI models tend to dominate our internal benchmarks.
It actually is showing in public benchmark if you know how to look for it. For example, in Terminal-Bench 2.1, GLM 5.2 received 78%, while GPT 5.6 Sol received 88%.
Then Terminal-Bench 3.0 came out (where the questions are new), and GPT 5.6 Sol received 34.6%, while GLM 5.2 dropped to a whopping 4.6%.
> OpenAI models tend to dominate our internal benchmarks.
That's odd, since Fable seems to be the leader for the industry. Not cost-effective, but if Anthropic models get dominated by OpenAI in your internal benchmarks, this calls their validity into question. Separately, see the jagged frontier effect. [1]
I've been using GLM 5.2 at my day job (mostly Rust backend work ATM). Nothing that blows away the models from OpenAI and Anthropic, but solidly good enough to get it done. A lot of people have experienced this and the fact that an open weights model can do so is where most of the excitement comes from. Optimizing for benchmarks can only get you so far, and people are quick to criticize models that fall into it (like DeepSeek Pro V4 recently).
That's interesting, we've had the same results as you at my company. gpt-5-mini was the clear pareto frontier leader for our in-house LLM benchmarks -- benchmarks that we built and tailored to our specific use cases. Then gpt-5.6-luna came along with the price cuts and immediately supplanted gpt-5-mini.
I always thought it was a little odd that gpt-5-mini was the leader for so long when more popular benchmarks placed gpt-5-mini further down the roster, but it seems you had the same result too.
> I would say GT is hardest to get in, but most inexpensive. While the later two are much easier.
Regarding the acceptance rates, it's actually the opposite. GT OMSCS has a very high acceptance rate (~80%) and UIUC and UT Austin online masters are much more selective (~30%).
The fact that Gemini returns the highest rate of fake positives aligns with my experience using the Gemini models. I use ChatGPT, Claude and Gemini regularly and Gemini is clearly the most sycophantic of the three. If I ask those three models to evaluate something or estimate odds of success, Gemini always comes back with the rosiest outlook.
I had been searching for a good benchmark that provided some empirical evidence of this sycophancy, but I hadn't found much. Measuring false positives when you ask the model to complete a detection related task may be a good way of doing that.
Sharding can be made mostly transparent, but it's not purely a DB-level concern in practice. Once data is split across nodes, join patterns, cross-shard transactions, global uniqueness, certain keys hit with a lot of traffic, etc matter a lot. Even if partitioning handles routing, the application's query patterns and its consistency/latency requirements can still force application-level changes.
> Once data is split across nodes, join patterns, cross-shard transactions, global uniqueness, certain keys hit with a lot of traffic
If you're having trouble there then a proxy "layer" between your application and the sharded database makes sense, meaning your application still keeps its naieve understanding of the data (as it should) and the proxy/database access layer handles that messiness... shirley
When sharded, anything crossing a shard boundary becomes non-transactional.
Ie. if you shard by userId, then a "share" feature which allows a user to share data with another user by having a "SharedDocuments" table cannot be consistent.
That in turn means you're probably going to have to rewrite the application to handle cases like a shared document having one or other user attached to it disappear or reappear. There are loads of bugs that can happen with weak consistency like this, and at scale every very rare bug is going to happen and need dealing with.
Two-phase commit provides an eventual consistency guarantee only....
Other clients (readers) have to be able to deal with inconsistencies in the meantime.
Also, 2PC in postgres is incompatible with temporary tables, which rules out use with longrunning batch analysis jobs which might use temporary tables for intermediate work and then save results. Eg. "We want to send this marketing campaign to the top 10% of users" doesn't work with the naive approach.
Sorry, what am I missing here, this complaint is true for all architectures, because the readers are always going to be out of sync with the state in the database until they do another read.
The nanosecond that the system has the concept of readers and writers being different processes/people/whatever it has multiple copies, the one held by the database, and the copies held by the readers when they last read.
It does not matter if there is a single DB lock, or a multi shared distributed lock.
These are limitations in the current PostgreSQL implementation. It's quite possible to have consistent commits and snapshots across sharded databases. Hopefully some day in PostgreSQL too.
I also had a close family member who committed suicide shortly after going on Prozac -- this also happened nearly 30 years ago. His young son later went on Prozac himself (several months after his fathers suicide) and immediately started demonstrating bizarre disinhibited anti-social behavior (e.g., damaging property, stealing from friends, etc). He was immediately yanked off Prozac when he started articulating his own thoughts of suicide. The bizarre anti-social behavior improved after discontinuing Prozac.
For some people, Prozac is a very dangerous drug. It is fully deserving of its FDA black label warning (which it didn't have 30 years ago).
That street view was filmed taken two months ago, when it was still warm and nice out, so tent activity would've been at its peak.
I wish I had seen your post earlier because I literally walked that stretch of road earlier this evening -- a couple friends of mine from out of town are going to the Patti Smith show at the Paramount tonight and we had drinks nearby. No tents in sight though we did encounter someone walking around with a blanket wrapped around their head. But still, one probable homeless drug addict is hardly "overrun".
It's really not as you describe... I agree things were getting worse in ~2019 and then became way worse during the pandemic, but it's much different now.
In the interest of saving myself an hour of time uploading a video, I’ll attest that yes—that street view is as “average day on Pike” as it gets.
To be clear, there are homeless who walk around the area… and Capitol Hill isn’t exactly the nicest area these days. 3rd and Pike isn’t nice. But Seattle in 2025 isn’t real-life World War Z.
It sounds like your last visit was during the COVID-19 pandemic. Homeless encampment conditions in downtown Seattle and throughout the city have much improved since then. Today, visible homelessness is effectively the same as it was back in ~2005.
I don't know why you're being downvoted. I've lived in the area for nearly 20 years and I agree that his description is far exaggerated. It was true in the depths of the COVID-19 pandemic (when any eviction or forcible displacement of people was prohibited), but not since then. Today, the visibility of homeless encampments in Seattle is roughly the same as it was 20 years ago.
> The homeless encampment problem in Seattle today is roughly now back at the same level of problem we had 20 years ago.
See, this is wild to me. Seattle has a long-time notorious problem with Tent Cities and even now it's just completely normal. I remember The Jungle under I-5 at Beacon Hill was a big problem. I'm sure it still is.
"The Jungle" isn't at all the same now, and is mostly cleared out from its heyday in ca 2018.
The larger problem with "The Jungle" is that nobody can agree on what it means, and where it is. I used to live on Beacon Hill, and the way journalists used the phrase was all over the place. Incidents anywhere from the I-90/I-5 interchange to the camps under I-5 all the way south to Georgetown, to the camps up in the woods in the East Duwamish greenbelt were all called "The Jungle".
You are citing the US News "best states" ranking. In that ranking, California is ranked #37 overall and Arkansas is ranked #47 overall. Even your own hand picked data source supports the OP...
37th/50 isn't good. But people never clamor on about how awful California is every time it's mentioned(well, rarely). This same ranking puts states like South Dakota and Indiana ahead, which I'm sure many would object to all the same.
This is the second day in a row I've watched threads about Arkansas of all places devolve into these nasty generalities(yesterday's was about WalMart and Bentonville). I don't live in Arkansas or anything, but I think we as a community can do better than devolve into it over and over, unless the topic at hand was the problems of a state.
I'm not saying that 37 out of 50 is good. I'm saying that 47 out of 50 is bad. Your data source doesn't refute the OPs argument that Arkansas is not a great place to live -- it actually supports the OP.
"But people never clamor on about how awful California is every time it's mentioned(well, rarely)."
Well, here's why it sucks, just so you can feel better.
1. the apartment rental industry is outright violating a court order that basically sets late fees and needs to outright face criminal charges. See Orozco v Casimiro
2. Ill-advised farmers up north are consuming tons of water on crops that make zero sense, and in many cases are growing crops that are basically exporting our scarce water overseas. Then they want to complain about a 'government-created dust bowl' when it's their own out of control water usage basically creating and exacerbating the situation. Oh and the desire to live in a floodplain, thus draining the largest lake in the state (if not landlocked lake in the country) and worsening drought conditions for the area for the past near-century.
3. OTHER STATES keep shipping their homeless here, thus drastically increasing our poverty numbers, artificially, and straining resources we don't really have thanks to ill-advised programs that do nothing to actually address anything.
4. Thanks to climate change, it's getting fucking HOT. Like, the heat you would normally only experience in the desert, is now a regular occurrence here in the valleys of SoCal. The deserts are actually cooler on occasion.
5. People can't drive and the cops do nothing even when it happens right in front of them. Oh, speaking of police, did you know many of them are in inter-department gangs? Yea, we just had to outlaw that.
6. Nobody's fixing the infrastructure. Sure new stuff is being built but that is supposed to be ON TOP of what we already have - and we're just letting what we already have crumble away. Yea internet and some power is going down but that's about it. Roadways, bridges, oh no. Route 66? It's screwed right now. Recent issues we had on the 40 forced a traffic reroute and the weight limit on those little bridges is 3 tons - guess what went over those bridges? Semi trucks with 10 or more tons of freight. You bet those bridges got wrecked, and nobody's fixing them. Ludlow to Cadiz is absolutely wrecked. Thankfully, I have an offroad vehicle and powerline roads exist, so I'm able to still get to digging areas or visit Dish Hill Volcano.
7. Thanks to new law, fast food workers have a minimum wage over the state minimum wage. So many jobs which require high skill, like what I do in LASERs and LED lighting, get paid less than them (I'm lucky, where I work knows my worth) and it ends up being demoralizing. I'm betting it caused a small hit across a few sectors as people said "I'll pay the $30 to get a license that lets me make $20 an hour" meanwhile starting techs in my sector get $16, or $17 on a night shift differential.
Happy yet? I can keep going. Rabbit hole's deep af.
Thanks, I think. But no, I didn't mean to pick on California as a bad state, I thought for some reason most people thought it was a nice state. I do, at least, despite its warts.
All I meant to say is that we can find many reasons every state is bad. But we shouldn't post them every time it's mentioned.
If Meta declared they were opening a huge new office in the bay, we'd get interesting discussion. If they announced they're opening it in Little Rock, we get little more than how awful AR is.
> Nobody's fixing the infrastructure. Sure new stuff is being built but that is supposed to be ON TOP of what we already have - and we're just letting what we already have crumble away. Yea internet and some power is going down but that's about it. Roadways, bridges, oh no. Route 66? It's screwed right now. Recent issues we had on the 40 forced a traffic reroute and the weight limit on those little bridges is 3 tons - guess what went over those bridges? Semi trucks with 10 or more tons of freight. You bet those bridges got wrecked, and nobody's fixing them. Ludlow to Cadiz is absolutely wrecked. Thankfully, I have an offroad vehicle and powerline roads exist, so I'm able to still get to digging areas or visit Dish Hill Volcano.
I feel like it's extremely misleading to give this example without providing surrounding context.
"Route 66" in this area is a 70 mile, parallel road to I-40 that serves a population of zero or nearly so and is a 2 lane strip of asphalt through the desert left over from before I-40 was constructed. It serves basically no function and no population, today.
It's not very obvious at all that it makes sense to spend money replacing the 100+ timber bridges along the stretch rather than just abandoning most of the road/downgrading it to a 4WD road with no bridges - although repairing/rebuilding it does seem to be what San Bernadino County hopes to eventually do.
Many of the bridges, while partially failing due to age, are also failing due to flooding damage - which is what caused the 3 ton weight restrictions to go in place in the first place and the sections that have been closed since 2014/2017.
-------
Additionally, the road is only officially closed east of Kelbaker Rd, and that section of road has zero population, and there's basically zero regional significance to the closure beyond mildly inconveniencing a few people in Amboy who can now only go West to get to I-40/civilization. (Especially since most reports I've seen suggest you can continue to the unpaved/rarely traveled Cadiz Rd anyway, which was the only connection that could only be accessed from the closed section).
The road from Ludlow to the Dish Hill Volcano is open, just less convenient if you used to get to it from the East.
""Route 66" in this area is a 70 mile, parallel road to I-40 that serves a population of zero or nearly so and is a 2 lane strip of asphalt through the desert left over from before I-40 was constructed. It serves basically no function and no population, today."
From end to end, Route 66 spans 2,448 miles.
"Additionally, the road is only officially closed east of Kelbaker Rd, and that section of road has zero population,"
The whole farm community in Cadiz would like to know they don't exist. I talk to the population there regularly before I go out to the Chambless skarn to dig. ditto the mining community that's there for the quarry at Kelbaker road (I own the uranium mine nearby.)
"The road to Dish Hill"
Collapsed last weekend at the railroad track crossover. You have to come from Amboy's power line backroads.
Great, but it's not contiguous. This stretch is about 70 miles long and merges back into I-40 at either end, and is the part of it we are talking about. What I said is accurate.
> The whole farm community in Cadiz would like to know they don't exist.
The farm in the middle of the desert that is a front for a decades-long attempt to loot and export the aquifer underneath it, regardless of the permanent damage it will cause, that one? That explicitly shouldn't exist.
It's also about 9 trailers and a house or two. And still has road access, as I noted and you appear to agree.
> Collapsed last weekend at the railroad track crossover.
Meaning the actual paved road to get to the area, or the unmaintained (unofficial?) path from the road, under the tracks, to the hill itself?
-------
But this is getting into the debate weeds. I'll be very generous and call the population along this stretch of road about 50 people.
There's close to 100 timber bridges in need of replacement, 70 miles of degrading asphalt, and of course - perpetual maintenance costs for both over the long term. It serves basically zero transportation function today.
I'm not a small government type, but the many, many millions of dollars it would cost to actually repair this road (not to mention continue to maintain it), do not seem remotely justified by it's utility.
The very tiny populations, tiny amount of industry, and very limited north-south function (Amboy Rd) here can be served by keeping the ~10mi Amboy-Chambless stretch and abandoning the rest of it/downgrading it to a 4WD track across the desert just like the dozens of other roads. Maintained/paved access via I-40 from Kelbaker Rd only.
-------
tl;dr - This is not, in my view, a case of government being unable to maintain infrastructure. This is a case where a large portion of the population would not feel that putting money into keeping this road in existence as a paved/2WD road along it's full length, is a worthwhile endeavor. You clearly have a vested interest in it.
I've assumed the same as well.
I also assume that many of the companies developing these models engage in benchmaxxing.
At my company we've developed our own internal benchmarks for evaluating LLM models as they become available. The benchmarks are tailored to our particular use cases but the utility and knowledge our benchmarks assess is still fairly universally applicable. I see wide differences between what our internal benchmarks report and what the major benchmarks do.
There was a whole lot of fanfare about how amazing GLM 5.2 was when it was released, but it was pure rubbish on our internal benchmark -- far behind OpenAI, Anthropic, Gemini, DeepSeek, etc. I don't know how to reconcile the fact that GLM 5.2 performed very well on some of the major public benchmarks, but consistently performs so poorly on ours. OpenAI models tend to dominate our internal benchmarks.