This is where I believe we are headed as well. Frontier models "curate" and provide guardrails, very fast and competent agents do the work at incredibly high throughput. Once frontier hits cracks the "taste" barrier and context is wide enough, even this level of delivery + intelligence will be sufficient to implement the work.
Taste is why I switched from GLM-4.6 to Sonnet. I found myself asking Sonnet to make the code more elegant constantly and then after the 4th time of doing that laughed at the absurdity and just switched models.
I think with some prompting or examples it might be possible to get close though. At any rate 1k TPS is hard to beat!
It was a little while ago but, GLM's code was generally about twice as long, and about 30% less readable than Sonnet's even at the same length.
I was able to improve this with prompting and examples but... at some point I realized, I would prefer the simplicity of using the real thing.
I had been using GLM in Claude code with Claude code router, because while you can just change the API endpoint, the web search function doesn't work, and neither does image recognition.
Maybe that's different now, or maybe that's because I was on the light plan, but that was my experience.
Claude code router allowed me to Frankenstein this, so that it was using Gemini for search and vision instead of GLM. Except that turns out that Gemini also sucks at search for some reason, so I ended up just making my own proxy which uses actual Google instead.
But yeah at some point I realized the Rube Goldberg machine was giving me more headaches than its solved. (It was also way slower than the real thing.) So I paid the additional $18 or whatever to just get rid of it.
That being said I did just buy the GLM year for $25 because $2/month is hard to beat. But I keep getting rate limited, so I'm not sure what to actually use it for!
No no! It was just the way you wrote it; but I think I misunderstood it.
> I found myself asking Sonnet [...] after the 4th time of doing that [...] just switched models.
I thought you meant Sonnet results were laughable, so you decided to switch to GLM.
I tried GLM 4.6 last week via OpenCode but found it lacking when compared to Sonnet 4.5. I still need to test 4.7, but from the benchmarks and users opinions, it seems that it's not a huge improvement though.
Last week I got access to Claude Max 20x via work, so I've using Opus 4.5 exclusively and it's a beast. Better than GPT 5.2 codex and Gemini 3 Pro IME (I tested both via OpenCode).
I also got this cheap promo GLM subscription. I hope they get ahead of the competition, their prices are great.
We reached some amount of early product-market fit and the “starter” stack that served us well early on: Vercel functions, Supabase Postgres, and a mix of queue services, started hitting limits. Function runtimes capped at 15 min, cron jobs had to be public, queues kept sun-setting, and the database had no automatic fail-over. We were spending more hours nursing infra than building features.
Over the next six months we rebuilt on plain AWS: EKS for the Next.js app, Aurora Postgres in-VPC, SQS for jobs, all managed with Terraform. A Cloudflare Worker mirrored an increasing slice of live traffic to Kubernetes until everything looked solid, then a DNS flip made it the source of truth.
Supabase data flowed into Aurora via logical replication; once lag was near zero we swapped the connection string and retired the old stack without users noticing. Costs are predictable now and we control most knobs.
We put this post together as a bit of a retro and to help if you're evaluating a similar path for your company.
We released an update to our pricing at Loops and along with increasing free limits, we made transactional email sending available at no additional cost on any paid plan.
We removed our previous plan tier of $1 per 2,000 sends.
Example:
You have 4,000 users and you send an automated onboarding email along with an occasional product update every month or two.
Your product also generates daily rollup emails, along with product notifications, and other standard transactional emails, maybe 100,000 emails a month total.
That's now available at the same price, just $49.
Say next month you gain another 500 users and you send 200,000 emails. That's still just $49 a month.
Our goal is to help you send all your emails from one platform for a single price point and this gets us one step closer.
We recently did some work improving our email deliverability to large institutions that use Rspamd as part of their corporate firewall, this is our write-up on that experience.
Super polished looking product -- any success in getting companies to use it? Work at a dinosaur that would never think of abandoning the microsoft ecosystem; wondering what the target is for something like Loops.
(I did try taking a look through the website and saw integrations posted more prominently than users?)
>the IPCC Sixth Assessment Report states "It is very unlikely that gas clathrates (mostly methane) in deeper terrestrial permafrost and subsea clathrates will lead to a detectable departure from the emissions trajectory during this century".
I do think that the potential behind GPT-4V is a bit understated right now. Obviously, roasting emails is not the best use case but there are some really exciting ones I'm interested in seeing other folks build.
reply