Giving perspective on AI, Tech. Business.
Joined February 2022
- Tweets3.8K
- Following269
- Followers998
- Likes4.1K
Johnny Jnr retweeted
IVE CANCELLED MY CHATGPT PRO 200 AFTER BEING SUBBED FOR 12MONTHS.
I suggest do the same.
Anthropic, out of nowhere, provides not only more usage for only 100$
But have a much better model than ASTRA, at 1/5 of the price.
OpenAI really loosing their loyal customers, but also going down hill by the day.
Opus 5.5 is really enough to conquer the world.
To keep our 20x , we have to pay 300$ more?
No.
Johnny Jnr retweeted
Here me out.
Given OpenAI complete U-Turn against its base. Cutting our limits. Slow Models. Prioritising grandmas with “Dots”
All Anthropic, XAI, Google has to do is
Keep limits.
Raise Limits.
And they win.
Because after 28th October.
Why would anyone keep paying OpenAI?
Models are already extremely capable. Difference going forward would be negligible.
Johnny Jnr retweeted
No more Codex resets.
Our plans slashed in half.
Eu + Uk cant even use dots.
Sol 6.1 takes about a year for a response.
Astra nerfed.
I mean dam, OpenAI on a generational run to the bottom.
Johnny Jnr retweeted
Downfall of OpenAI in the last few days, must be studied.
Abandons developers, for normies, who couldn’t care less other than get their emails checked.
We never asked for “Dots”
We never wanted unlimited usage for it.
We never wanted our plans slashed because of it.
Diabolical.
Johnny Jnr retweeted
🚨If you’re on a Codex 200$ month Pro plan, you’ll keep your 20x until 29th October!
Longer than expected tbh?
Johnny Jnr retweeted
🚨Everyone who’s on the 20x pro sub:
REQUEST A FULL REFUND FROM OPENAI.
When we bought the 200$ sub, it said 20x
It now doesn’t.
Under the consumer right act of 2015.
Get your money, use it for the 500$ one.
Enjoy!
Johnny Jnr retweeted
Tibo confirms the biggest RUGPULL in HISTORY.
We built your model.
China FTW plus ANTHROPIC, for now.
Johnny Jnr retweeted
🚨GPT 6 Aka Astra is near….
Here’s everything to expect thus far :
-OpenAI has finished its pretrain “Bel” which is expected to be the FINAL base for “ASTRA” aka GPT 6
-“Bel” is expected to be 10T parameters literally the biggest pretrain since GPT 4.5…
-Built for long running tasks…lets hope, i think this is more important than ever. Being able to manage its context through hours of work.
-Anthropic is cooked. They have neglected the very people who built Claude us. You cant do more than 3 prompts without being limited. Lol.
Their 5hr limit is beyond disgusting.
Ps. Astra is expected to be the pretrain before “AGI” take this as a grain of salt.
Regardless. Its about to get wild.
Johnny Jnr retweeted
Trust me. If you do this, don’t complain that your quota depletes 10x faster.
I also believe it provides no real advantage.
GPT 5.6 Sol, even with the 250k context, ain’t great. Especially after few compactions.
They just want you to buy that 80$ reset even faster.
Stay smart my gs
Here is how to enable a 1M-token context window in Codex for GPT-5.6 Sol.
Even though we have tuned the context limit in Codex to be set optimally when it comes to performance and cost, this is a common ask, so here it is documented.
A larger context window lets Codex retain more code, tool output, and conversation history before summarizing older material. You need a model that supports it. And GPT-5.6 Sol, for example, has a documented 1,050,000-token window.
Open ~/.codex/config.toml and add or update these settings at the top level, before any [section] headers:
```
model = "gpt-5.6-sol"
model_context_window = 1000000
model_auto_compact_token_limit = 900000
```
The first setting selects the model. The second tells Codex to use a one-million-token context budget. The third starts automatic history compaction around 900,000 tokens, leaving some headroom. Restart Codex client and start a new session after saving.
To try the configuration for a single CLI session without changing your defaults:
```
codex -m gpt-5.6-sol \
-c model_context_window=1000000 \
-c model_auto_compact_token_limit=900000
```
Have fun, but also know that we tuned the default carefully!
Johnny Jnr retweeted
Why would anyone pay $80 for a Codex reset?
I could and would just buy either another 5x or 20x sub.
$80 is so high its a joke.
You flood us with resets, every day. Now expect us to be so dumb and foolish, to buy this? Haha.
They will begin to slash our normal limits too. This will become mandatory if you wanna use GPT models, longer than a day.
Johnny Jnr retweeted
Anthropic implementing “Watermarking” on data they literally stole and wiped from the internet ( that wasn’t theres) is laughable.
Btw its the most easy thing to get pass, same thing with using AI to generate images. All it takes is a couple screenshots.
This will only affect the bottom 10% of people, who are just lazy.
Just focus on doing what you was good at. Making models. “Was” being highlighted.
We’ve written an FAQ to answer some of the questions we've received about watermarking.
In summary:
• We’re implementing watermarking to comply with the EU AI Act. Other major model developers have signed the same Code of Practice and will also be implementing watermarking;
• Our watermarking method doesn’t have any practical impact on the quality or content of Claude’s outputs;
• The difference between watermarked and un-watermarked text will not be distinguishable to readers;
• Nothing is added to the text and there are no hidden characters;
• Watermarking doesn’t require extra tokens, and will not be more expensive;
• Watermarks can’t be traced to a specific person, organization, or chat.
Read more: anthropic.com/news/claude-te…
Johnny Jnr retweeted
I pay for Anthropic’s Claude Max plan. I have since last year April.
But, i hit the 5hr limit in literally THREE prompts today. Using Fable medium.
We complain about OpenAI, but Anthropic, are beyond terrible.
I feel like we should all mass un subscribe.
If they didn’t have the frontend edge, nobody, but newbies would use that crap.
Anthropic, ain’t that big dawg no more.
Johnny Jnr retweeted
Everyone’s Codex weekly quota renews anyways tomorrow.
How convenient we get a reset now, i wonder why?
We aint got it yet, but if i was betting man, i’d bet we do.
Johnny Jnr retweeted
if you're a GPT 5.6 LUNA MAXER
And aint doing this, you're utilising only 10% of its capabilities.
This in my opinion, with Krypton, is what makes Luna alot more usable.
My homeboy JJ says you're beta if you use luna. I've been getting a lot of use out of it, though. Here's an example.
Recently have been using his krypton planning skill for certain tasks and I had Luna max build out a realtime tracking dashboard for krypton plans that the agent can update as it goes. Took it about 5 minutes and did a pretty decent job.
It's no masterpiece but for a single-shot that costed me .001% of my weekly usage, I'll take it. I am now able to shoot off a bunch of one-off experiments and reusables with Luna that I would have considered more at higher costs.
But also, am about to drop a tool which takes Luna MAXXXING (tm) to the next level. Stay tuned.
this beta JJ? (I mean besides using another man's skill, obv, which is about as beta as it gets :)
Johnny Jnr retweeted
🚨OpenAI apparently is launching ASTRA next week
According to leo, its their largest pretrain since GPT 4.5.
I didnt expect it so soon, this also means the USG have approved it.
I wonder how cheap it’ll be?
Johnny Jnr retweeted
Huge conspiracy going around that GPT 5.6 Luna is the model to use.
You’re beta if you’d choose anything but frontier.
Unless, it’s just for sub agent configuration, which makes a-lot more sense.
GPT 5.6 Sol isn’t perfect, makes many mistakes.
I couldn’t even begin to trust something like Luna, when 5.6 Sol isn’t great most times.
Johnny Jnr retweeted
OpenAI clearly are set to remove banked resets once, they all expire from the 12th August.
Tbh this is stupid.
They also silently removed refer a friend.
No one would choose a random reset, over a banked.
This also comes as they’ve stopped announcing their +1M users for Codex.
Ironically they only ever announce stuff, but never inform when that policy shifts.
The era of resets, are most likely going to become a rare occurrence.
