There was question about how live meeting assistant works. 1/ you have to use the browser version of the meeting (not native apps) and have the chatpanel open while the meeting happening. you can ask any question about the meeting while the meeting happening and chatpanel will use configured AI agent/llm provider to answer it. Towards the meeting end it can summarize. the best part is you can ask all sorts of questions after meeting ends which is missing from out of the box meeting features provided by the meeting provider. It depends on the closed captions being enabled and capture that as live scribe. I am planning to add browser audio caption feature as well in the future
The worst part of Opus that I dont like is they control what you can/can't do. the guardrails that they do in the name of interpretability where you steer you. Last couple of days I was working a project that supports bunch of models and the model said only Claude can do it start writing code that doesn't work with codex, opencode, pi etc. Finally, when I switched to Codex, everything worked. To me, they are controlling the narrative. This is Opus 4.8 vs GPT 5.5.
I can't believe I would say this. I TRUST OpenAI more than Anthropic. They try to play best actor but they are manipulating the behavior of the model in the name of guardrails/interpretability.
That is why I refuse to build anything that works with Anthropic models as the backend. Because, when they want to shut you off, they can do it by just making model less reliable in your product than their offering!
And it's really good and fast. Have tested with bunch of odd photos on what is happening. Overall the training set seems large enough to know what's what and where
You can one-shot a port of Linux to Rust and stop contributing to open source.
The value of software is going to tend towards zero. The value of the software developer the same.
Anthropic is now a kingmaker. It gets to decide which businesses get the expensive private model that can generate entire business functions at the drop of a hat. If you can't afford the price tag, then competition in the market is not for you.
Computing is no longer "personal". It's for big biz only.
GP is exaggerating but I am convinced this will happen sooner rather than later. The improvements in AI are truly exponential if you read the SOTA papers. It's hard to keep up week to week.
as someone who uses these models day in out, i can confidently say its more of a marketing gimmick than anything else. don't get me wrong, the model is great, but nits no out of the world than GPT 5.5 or similar ones. I would say just go and try this model for serious work and see the marginal difference. the model wins in some cases and loses in many others. so, what is this all about? hype!
Working on my codebase (~100KLoC across multiple Python modules) I felt that Fable was head and shoulders above 4.x series. It was just relentless and always hell bent on testing and proving its own work. It just tore through problems like an animal. I never seen that behaviour in 4.5-4.8. I can't speak for OpenAI models as I don't use them but Fable was in a different league. Especially when tasked with long horizon goals that involved reasoning at a high and low level to solve the task.
I think a lot of users likely use these models on small hobby projects and not some convoluted enterprise code base. When you're making yet another Space Invaders clone it really won't show much difference. Messy, complex code bases with layers of cruft from decades of patching - that's what separates the model boys from men.
Yeah, and its browser usage on tough web apps/sites was also amazing. This is one of the cases where it is easy to tell a difference. It was figuring out very effectively how to find right elements whereas with previous LLMs I had to constantly babysit and unblock them with browser usage.
I used codex 5.5 and Claude. I pay for Claude from my pocket. I use Codex at work. I can confidently say Codex 5.5 high is much better in going through long code bases (couple of millions of lines of code) vs Claude Fable/Opus which does only what is been told. while codex covers all sorts of edge cases. Frankly, I am not going to miss a thing if they stopped Fable.
That's raw NASA SDO satellite footage. Claude (Opus 4.7) was used almost exclusively for building the site. Static site on Render (no hosting fees), pushed from Github. Uses NASA API's (free), a very cost-friendly project on the ole wallet!
I'll add that "raw" is after a bit of postprocessing to make it pretty.
When the SDO webserver went down a few months ago I rebuilt the L1 data processing pipeline from JSOC so we could still do outreach and there's a surprising amount of opinion that goes into the mapping of data to visualization for each wavelength. My composite movies came out looking more like an acid trip than solar data.
Touché — when the person who rebuilt the pipeline says it's not raw, it's not raw :)
Is optical-flow interpolation a step too far for outreach, or fair game? Tempted to motion-interpolate (ffmpeg's minterpolate) the daily MP4s up to 60fps for Lumara— looks gorgeous but the in-between frames are extrapolated. You're totally right about "raw", I suppose I meant more straight from NASA APIs.
I personally would rather show actual data over interpolated data, but I don't know how many unphysical interpolation artifacts you're getting or whether that really matters at a public outreach level.
That said, if you felt like processing it yourself, the L1 files are available every 12-24 seconds, and preprocessed images are available ~1 per minute. The synoptic version you're using is about 1 frame every 3 minutes at 20 fps so you could just triple the framerate without needing any interpolation, or string together the images yourself.
I am actually super impressed with Codex-5.3 extra high reasoning. Its a drop in replacement (infact better than Claude Opus 4.6. lately claude being super verbose going in circles in getting things resolved). I stopped using claude mostly and having a blast with Codex 5.3. looking forward to 5.4 in codex.
Same, it also helps that it's way cheaper than Opus in VSCode Copilot, where OpenAI models are counted as 1x requests while Opus is 3x, for similar performance (no doubt Microsoft is subsidizing OpenAI models due to their partnership).
I've been using both Opus 4.6 and Codex 5.3 in VSCode's Copilot and while Opus is indeed 3x and Codex is 1x, that doesn't seem to matter as Opus is willing to go work in the background for like an hour for 3 credits, whereas Codex asks you whether to continue every few lines of code it changes, quickly eating way more credits than Opus. In fact Opus in Copilot is probably underpriced, as it can definitely work for an hour with just those 12 cents of cost. Which I'm not sure you get anywhere else at such a low price.
Update: I don't know why I can't reply to your reply, so I'll just update this. I have tried many times to give it a big todo list and told it to do it all. But I've never gotten it to actually work on it all and instead after the first task is complete it always asks if it should move onto the next task. In fact, I always tell it not to ask me and yet it still does. So unless I need to do very specific prompt engineering, that does not seem to work for me.
That shouldn't really make a difference because you can just prompt Codex to behave the same way, having it load a big list of todo items perhaps from a markdown file and asking it to iterate until it's finished without asking for confirmation, and that'll still cost 1x over Opus' 3x.
The only reason I was sticking to Android for years is this. And I think there is no moat for Android. I would rather switch to iOS if both platforms are same restrictive.
I did this last year. Reluctantly. And using iOS still hurts. But it’s better than that Google crap.
I developed my own Android ROMs from 2009-2011, complete with my own tuned kernel. I ran the local Android developers MeetUp group and evangelised Android development. When Honeycomb launched I helped OEMs test their beta firmware. For free.
But as Google has become certified Evil, the direction of Android has been very clear. In practice I honestly can’t say it’s now any more open than iOS. Except it has a lot more avenues for Google to mine your data to sell ads. And the quality of third party apps on it is decidedly worse.
I thought long and hard about getting a Linux phone. But I need a good camera on my phone to take random snaps of kids/pets/etc. And the Linux phones just aren’t there.
I hate the shitty duopoly we have ended up with. But I now realise that the openness of x86 and pc as platform really was an accident of history.
When you say quorum what do you mean? Is it like an agent swarm or using all of them in your workflow and independently they perform better than opus? Curious how you use (tooling and purpose - coding?)
reply