Posts on X before Apr 7, 2026, 8:56:03 PM UTC
I mean, we all understood this was going to happen right?
they've finally created the seed of a psuedo-god that does their bidding, literally grants them unlimited wishes in anything near proximity and within the virtual world, ...
and you think they will sell it to you ASAP?
by that time they'll have an npm package called 'mythos-patch', you run it, it oneshots your entire repo/monorepo into impenetrability
Okay so Anthropic's Claude Mythos is so good they are sharing it with big companies so they can patch vulnerabilities before releasing it to general public? Who will give me my 15 days to safely fix vulnerabilities in my apps which Mythos will expose to everyone? Chat, are we c… Less
Okay so Anthropic's Claude Mythos is so good they are sharing it with big companies so they can patch vulnerabilities before releasing it to general public? Who will give me my 15 days to safely fix vulnerabilities in my apps which Mythos will expose to everyone? Chat, are we cooked?
gpt 4.5 is a bad example here because it was useless other than for some half decent prose upside for slopwriters
again, if this mythos model measures up to even half of what is being shown it's capable of, price does not matter here
Anthropic: "We do not plan to make Claude Mythos Preview generally available" A big line, buried quite deep. Possible reasons? So many, inc: 1) The model is expensive (25/125), not far off GPT 4.5, which became commercially unviable. Less likely, given the claims about Mythos… Less
Anthropic: "We do not plan to make Claude Mythos Preview generally available" A big line, buried quite deep. Possible reasons? So many, inc: 1) The model is expensive (25/125), not far off GPT 4.5, which became commercially unviable. Less likely, given the claims about Mythos. 2) They genuinely are worried about unleashing cybersecurity choas on the world. 3) They don't have capacity to serve it at scale yet. 4) They will quickly distil the early access outputs of Mythos into a lighter model, so no need to release the bigger model when a more cost efficient one coming imminently. 5) Other. Not read the 250 page report yet, but will do.
I love this joke every time a new model comes around with "muh security!!!" essays and ID verification walls
but if you've been paying attention, this one is different
people already forgot about Opus original pricing
$25/$125 is more than welcome even if the benchmarks are half as good of an upgrade as they claim x.com/flavioAd/statu…
And just like that, everything is being ~solved in 2026
>Anthropic's Mythos ~solving code generation
>OpenAI's ImageGen v2 ~solving image generation
>Seedance 2 / Veo 4 ~solving video generation
>ElevenLabs v3 ~solving voice/effect generation
>Sora v5.5 ~solving music generation
So @TimJDillon could make a full podcast episode every single day instead of per week, and it STILL wouldn't be enough to cover what's been going on/what's going to happen next
so cool to see people finding out about modafinil/armodafinil, l-theanine, etc
i wonder how much longer until they find out about piracetam, oxiracetam, nootropics in general
really takes me back ~15 years ago when i was in high school buying all this stuff online lmao
it's nice to see trading going in this direction.
web + mobile, tradingview, market data, trade copiers, marketplace, multi-prop support, one login. feels like a much more complete direction for the space.
sean's post: linkedin.com/posts/seanrkoz…
if you're not spending your time upgrading your repos/your clients' repos to be less and less dependent on... dependencies, you're ngmi (nor will your clients)
every single package is an attack vector ESPECIALLY when today's agents are so trigger happy with updating to latest
TeamPCP has done ANOTHER supply chain attack. My Brother in Christ, how many of these fuckin' things are you going to do? YOU'VE DONE 50 FUCKING SUPPLY CHAIN ATTACKS. 50 SUPPLY CHAIN ATTACKS IN EIGHT FUCKING DAYS. March 19th: - Trivy March 20th: - EmilGroup (28 packages) - Ope… Less
TeamPCP has done ANOTHER supply chain attack. My Brother in Christ, how many of these fuckin' things are you going to do? YOU'VE DONE 50 FUCKING SUPPLY CHAIN ATTACKS. 50 SUPPLY CHAIN ATTACKS IN EIGHT FUCKING DAYS. March 19th: - Trivy March 20th: - EmilGroup (28 packages) - OpenGov (16 packages) - Teale-io (eslint-config) - AIRTM (uuid-base32) - PypeSteam (floating-ui-dom) March 23rd: - Checkmarx March 24th: - LiteLLM March 27th: - Telnyx
this is the state of google rn vs <1s in my terminal
can you find the answer on the entirety of the google page?
alright my prediction was for 2027 but it is increasingly looking like 2026 is the last year using dependencies is a good idea
if you're making serious software rn, the worst investment you can make is dependencies
should be obvious but: - anthropic is going all in on next model due to breakthrough; every bit of compute is precious; short term loss (lowering usage) for long term gain (quicker to the next level of models) - openai is going all in on the next model due to breakthrough; wipe… Less
should be obvious but:
- anthropic is going all in on next model due to breakthrough; every bit of compute is precious; short term loss (lowering usage) for long term gain (quicker to the next level of models)
- openai is going all in on the next model due to breakthrough; wipes sora completely out and uses absolutely everything towards the next gen of models
2026 is the takeoff even normies will feel, iykyk
to help you understand the current situation of AI, specifically LLMs, this is all you need to grasp really:
disney is just a lil bro to OpenAI
dependencies are a thing of the past soon imo
especially with what happened with litellm
maybe EOY 2027 it will be mainstream
the for you page is absolute garbage again @nikitabier
- WHY are my likes not weighed?
- WHY are my bookmarks not weighed?
you always take a look at people's accounts, right, so take a look at mine - my likes, bookmarks - and tell me, WHY does my For You look like it does?
for those saying that medium/high reasoning performs better than xhigh reasoning, it's easier to correct mistakes due to overthinking than to correct mistakes due to not being diligent enough in the research/plan process it almost feels like medium/high have a hidden prompt "ju… Less
for those saying that medium/high reasoning performs better than xhigh reasoning,
it's easier to correct mistakes due to overthinking than to correct mistakes due to not being diligent enough in the research/plan process
it almost feels like medium/high have a hidden prompt "just do this ASAP for the user, wing it". that's what those reasoning levels feel like in comparison
one of many things that keep me going and smiling at the start of every day:
- knowing that i haven't yet unlocked this skill tree (hardware)
- it gives me perspective on how non-software people view my work (by how I, a non-hardware person, view this type of work)
currently the only proven way is to first rip out entirely what you are wanting to refactor, then make a new session, and tell the new agent what you want to implement
now you also need to make sure that your memories have no mention of that ripped out feature as well
oh look, a dir that shows what companies to invest in because of how early they are to the latest bleeding edge tech of the future skills.sh/official
no one is sharing, it hasn't been able to keep up, and now that gap will widen exponentially
the best they can do is imitation of these models
systems complexity should come from the system, not the AI set up
put harnesses and blockades up for your runs, not your agents
pi was already so super fast, now it's blazing away with the parallel tool calls executing simultaneously, amazing stuff @badlogicgames
friendly reminder, Q1 2026 hasn't even ended yet
if you think the field is unrecognizable since December (which if you don't feel that, you're not pushing hard enough/up to date)...
just know it hasn't even been a full 3 months since then
imagine 9 months (EOY). lmao buckle up
soon, salary is going to matter a lot less than your ITB (individual token budget)
working for $250k with a $1m token budget per year will be leagues more fun/growing than even $500k but needing to beg for tokens
this is already the case at the top of the competition
the problem with this tweet is that if anyone else said this we'd all laugh
but it's elon we're talking about here
most will still laugh, sure, although that laughter is now nervous laughter rather than making-fun-of laughter
if you haven't realized the below statement after seeing what @karpathy and @tobi have done just this past week or so,
you're ngmi :(
it's the first model to refer to itself constantly as "model", as in,
"model thinks xyz. model thinks that's good idea and something to explore. model thinks this is a benchmark and will tryhard."
Rarely should you ever be steering with follow-up messages.
Most of the time, queuing next messages produces much higher-quality results than steering the model in the middle of what it's doing.
RAM on your macbook will basically become an intelligence potential marker
instead of "how many things at once can I be doing" it turns into "how much intelligence at once can I be generating"
Ultimate marketing move here @thsottiaux
reset Codex limits for the reason of Claude being down 😂
you are no longer the violinist in an orchestra,
you are the conductor
Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just looks very different because humans now move upwards and program at a higher level - the basic unit of interest is not one file but one agent. It’s still programming. … Less
Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just looks very different because humans now move upwards and program at a higher level - the basic unit of interest is not one file but one agent. It’s still programming. x.com/karpathy/statu…
bigger than you realize
Gemini Embedding 2: Our first natively multimodal embedding model
Gemini Embedding 2 is our first natively multimodal embedding model that maps text, images, video, audio and documents into a single embedding space, enabling multimodal retrieval and classification
one is the original intelligence - whatever that word means to you - while the other one is a copycat of that intelligence
poetic in a way if you think about it
i dont really see the endgame of AI labs hundreds of billions of dollars spent for a SOTA model just for a random open source model to come around and similar benchmarks doest it even matter to be “first” to whatever the goal is? or just surviving long enough i.e. Apple will … Less
i dont really see the endgame of AI labs hundreds of billions of dollars spent for a SOTA model just for a random open source model to come around and similar benchmarks doest it even matter to be “first” to whatever the goal is? or just surviving long enough i.e. Apple will simply just take what works without burning any cash
this has been said at least once a day somewhere around the world for the past 60+ years of CLIs
CLIs are not going anywhere lmao
damnit @bryan_johnson you were right about nothing beating 8 hours of sleep. no amount of caffeine (or anything else 😏) replaces it
I was doing 6 hours for most of my 20s, if not less
so much more done with so much more clarity and quality than before
GPT 5.4 xhigh is the first model to have a semblance of high quality writing, feels like some of 4.5 rubbed off on it
Still not jaw dropping but if you deal with a lot of text (non-code) you're sure to feel the difference
Customer Support -> Customer Engineer
The former becomes the latter in the next few years; typical support will be automated away while custom engineering/fine tuning will replace it.
gpt 5.4 xhigh is the first model that can change your life even if you start completely from scratch right this moment
every model has its favorite word
gpt 5.4's favorite word is "seam"
even if next year's open source models are as good as today's closed, the closed ones of next year will make using those open source models a waste of time
it's been like that for a while if you've been paying attention these past few years
I don't know who wants to hear this, but buying an expensive MacBook Pro or Mac Studio to run small LLM models is useless. These models have low density of world knowledge and reasoning. It's much cheaper and faster to use an API. Maybe next year they will be useful, but not this… Less
I don't know who wants to hear this, but buying an expensive MacBook Pro or Mac Studio to run small LLM models is useless. These models have low density of world knowledge and reasoning. It's much cheaper and faster to use an API. Maybe next year they will be useful, but not this year.
it is jaw dropping how terrible antigravity / gemini 3.1 pro is
works of programming genius like this, especially those made before pre-2000's, are always awe inspiring
someone out there did all of this, efficiently, no templates, youtube tutorials, AI (lmao!), wasn't mainstream at all especially in the 80s,
humans. we had a great run.
Some games had the rare quality of being entirely unique, introducing concepts that had never been done before. These days, most games are variations of ideas that have existed for a long time. One of those unique-concept games that has stuck with me over the years is Marble Mad… Less
Some games had the rare quality of being entirely unique, introducing concepts that had never been done before. These days, most games are variations of ideas that have existed for a long time. One of those unique-concept games that has stuck with me over the years is Marble Madness. I was instantly hooked the first time I played it. Although I wasn’t very good, the urge to return and try again was incredibly strong. It had an almost mesmerizing quality. The graphics - for 1984 - were impressive, as were the music and sound effects, which added to the “this feels very different” vibe. Most fascinating, though, were the physics: the marble reacting to slippery surfaces, bumping over ramps, and being knocked aside by the “evil marble.” It demanded real skill and patience to navigate the levels. The unique gameplay also makes it timeless in my opinion. Marble Madness looked great over 40 years ago and will still look and play great 40 years from now. The only downside was its relative shortness with only 6 levels. Mark Cerny originally wanted to create more but the memory limitations were the main reason for the small number of levels.
Following the upcoming special Security College on Monday, I am convening a special Zoom call in 3 months.
It is of the utmost importance that there is no further escalation until then.
If you haven't felt it yet, OpenAI has won.
Ever since their GPT-5 line of models, it's become more and more clear with each iteration they release.
btw, building your own transcription solution is something you should have done two years ago with Sonnet 3.5
let me finally all-in on Atlas and drop Chrome by making Atlas 100% controllable via Codex CLI
BONUS POINTS if I can continue to use Atlas/browse on my own (scrolling X) while Codex CLI + Atlas do stuff together like e2e testing or filling out forms or XYZ
