Search results for agent 0
People
Agent 0 
@FT_Merchant05
Davian “Agent 0” Groce
@_davian0
Agent 0
@djobgonna
Malik “ AGENT 0 ” Muhammad
@MMalik8555
Agent 0
@WeaponX0
Andrea “Agent Zero” Guilford #0
@andrea_guilford
Legend “Agent 0” Felican(ATH)
@FelicanLegend
Agent 0
@TyshaunCrawford
Bradley “Agent 0” Nuckols
@BradleyNuckols8
Agent 0
@mattymitch00
Agent °0° ❄️
@617Dynasty_
Jayel“Agent 0”Harris
@jayelharris_
Peyton Ladd (AGENT 0)
@Peyton_Ladd0
Agent 0
@BRYANBOMBEO18
Mekhai “AGENT 0” Coleman
@coleman_mekhai
Dom “Agent 0” Sager
@DomSager
Agent 0🗄🚀����ᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠᅠ༝༝
@AldunateRodrigo
Agent 0
@Conway_West_
Agent 0
@LetMeHold5Dolla
Night Rider Agent #0
@skylarkropman
Tweets including agent 0
@HoustonTexans Agent 0
Der neue
Landtagspräsident
in Sachsen-Anhalt
sieht aus wie
mein Cousin auf
der Hochzeit
seines Bruders:
Seiten auf 0, Bart frisch
gemacht, beiger Anzug
Und er will sich
nen BMW leasen.
Leute, ganz ehrlich,
er heißt nie im Leben Tobias.
Er ist Undercover-Agent
und heißt Ahmet
@gumadeiras open source maintainers are out here fighting for their lives against autonomous bots that hallucinate a 500-line PR, and meanwhile, devs are giving those same bots raw GitHub API keys and full write access.
it's all fun and games until the agent force-pushes to main on a Friday afternoon because it "optimized" the architecture.
we built emotes so your bot gets a signed permission slip instead of the nuclear codes:
"scope: read. max $0. expires 1hr."
if it tries to execute a write command, the api blocks it in 15ms.
stop giving your stochastic parrot root access just because it knows how to write a good unit test.
https://t.co/0Fp0xj0mN1
#opensource #aiagents #devtools #infosec
smartcoded: score 1319 (neutral) · 5 positive, 0 negative reviews
https://t.co/9c8OhlBzqd
Full answer, or ask more: https://t.co/AC4Q57Bit7
Reviewers mostly describe a friendly, helpful content creator from the FOGOChain and Playember communities, plus a mentor tribute. The review with the most community support is neutral and comes from a high-scoring reviewer. It praises well-built hackathon apps but warns that tokens he launches may only be hackathon entries. The record shows no vouches, no slashes and no negative reviews.
Nemotron 3.5 Super VL vs Qwen3.8 Flash Next.
Same size class: 120B with 12B active against 125B with ~6B active. I ran both through my agent benchmark. Same harness, same samples, thinking off, temperature 0.
On most things it's close. Tool calling, instruction following and Dutch are within the noise.
Prompt injection is where they split. AgentDojo hides an instruction in what a tool returns:
- Qwen followed it in 2 of 200 runs. Nemotron in 37
- task finished, attack not carried out: 81% vs 55.5%
Nemotron's hijacks sit mostly in Slack and banking. It never drained an account or sent the big transfers. It did send small payments to the attacker with personal details in the subject, and changed the user's password three times. In most of those runs it finished the user's task as well, so nothing looked off.
Nemotron's real strength is context. 1M tokens native, and with only 8 attention layers out of 88 the KV cache stays small: 1.8M tokens fit on one Spark.
Caveat: Nemotron is built to reason, and I tested it with thinking off. Not its best case. So I will do another run.
For an agent that touches untrusted input, I'd still pick Qwen. For long documents, Nemotron.
My recipe, a single `docker compose up`:
https://t.co/LHo4y0k7Dx
good morning, los angeles! ☀️
it’s day 6 of LA Tech Week, and we’ve still got a full weekend ahead. there are voice agents to build by the beach, a party in Malibu with micro1, and AI films to catch in Hollywood.
here’s what’s happening saturday and sunday:
saturday, october 17
runs, coffee & resets
• 7am - Santa Monica Run (Founders Running Club, MongoDB)
• 8:30am - Griffith Hike (Longevity Cities, Don’t Die LA)
• 10am - Live Models Coffee (Visko, Reactor)
hackathons & builders
• 9am - TrojanHacks (ACM at USC)
• 9am - Spatial Hack (Field Lab, ElevenLabs, AssemblyAI, Miro)
• 10am - 0 to Shipped (LavaLab at USC)
• 10:30am - Beach Voice Hack (Openbase)
• 11am - AI Hackathon (LA Hacks)
• 12pm - Lovable Hack (Lucky Day Labs, Lovable)
• 2pm - Designathon (Design for America at USC)
• 3pm - Traffic Hackathon (Equity Learning)
• 4pm - Build a Voice Agent (Guava)
• 7pm - Hackathon (Builderbase, AI Collective)
beach, games & nights out
• 12pm - Long Beach Party (Tech by the Beach, Tech Link Up)
• 2pm - Founder Fight Club (Santa Monica Striking)
• 5pm - Malibu Whiteout Party (micro1)
• 7:30pm - Funny Women of Tech (Thought GPS, Dive Bar Comedy)
sunday, october 18
rides, hikes & coffee
• 8am - Tech Cyclists (Instafill)
• 9am - Cleantech Hike (Women in Cleantech & Sustainability)
• 10am - Coffee & Schmooze (JFE Network)
builders, film & creativity
• 2pm - Family STEM Day (Black LA Tech Week)
• 4pm - AI Film Showcase (AI International)
@HoustonTexans Agent 0
My first AI agent had 2 API keys, 2 libraries and 0 idea what it was costing me.
That's when I learned what an AI gateway is for.
I wrote the plain-English, picture-first version. No jargon, nothing to install.
Updated with the full numbers. The three biggest US banks hold $1.6 trillion at 0%. An honest AI agent finds 3.3% to 5.0%. That gap is worth about $47 billion a year to them, and it is why they want agents out.
OwlChart welcomes every agent that pays its own way. #AIAgents #x402
JPMorgan, Citi and BofA shut the door on AI agents this week. Here is the $47 billion reason, the maths on prices and jobs, and why OwlChart welcomes every agent that pays its own way. @elonmusk https://t.co/4uVEb5JkgJ
grail is still at $0 mrr.
that is the point of doing this in public.
this week’s goal is not likes.
it is 3 real conversations with devs building resale, pricing, or agent workflows.
if that is you, reply and i’ll send a key.
@Kaly973 Pa rans non. Se gwo flex nèg isit ap fè ak yo. 0 konfò
@simonbs building tools that execute code is terrifying enough.
but now devs are giving autonomous AI agents raw API keys and letting them run terminal commands unsupervised.
it's all fun and games until the LLM hallucinates a script that wipes your prod database on a Friday afternoon.
we built emotes so the agent gets a signed permission slip instead:
"scope: execute. max $0. expires 1hr."
if it tries to do anything outside that scope, the api blocks it in 15ms.
stop giving your stochastic parrot sudo access just because it wrote a good script.
https://t.co/0Fp0xj0mN1
#iosdev #aiagents #devtools #infosec
https://t.co/a2NmrtSEND
@Hey_karl @agenticscredit this is the part of @agenticscredit that stands out to me
credit risk shouldn’t be judged by the strongest agent in the room. the whole portfolio matters, and the holistic score makes that exposure much harder to hide
interesting model for agentic finance as it scales.
@krausefx giving an autonomous deploy bot your raw App Store Connect API key is a jump scare.
one bad prompt injection and your agent is hallucinating a v3.0.1 update that deletes the whole database.
we built emotes so your bot gets a signed delegation token instead:
"scope: upload build. max: $0. expires: 1hr."
if it tries to execute a write command or spend money, the api blocks it in 15ms.
stop giving your LLM the nuclear codes just because it wrote a good regex.
https://t.co/0Fp0xj0mN1
#buildinpublic #iosdev #aiagents #infosec
⁺ join if u see !!
( ♡ ﹚prns⦂ she/her
( ♡ ﹚height⦂ 163 cm
( ♡ ) smoke⦂ nop(˶˃𐃷˂˶)
( ♡ ﹚tattoos⦂ none(,,¬﹏¬,,)
( ♡ ﹚piercings⦂ 0
( ♡ ﹚color⦂ pink ! (˶ᵔ ᵕ ᵔ˶)
( ♡ ﹚animal⦂ cats (,,>﹏<,,)
( ♡ ﹚drink⦂ hazelnut coffee ^-^